SU\G(𝔸)/K·U
Sign Up for free Log In

Author Description

Login to generate an author description

Ask a Question About This Mathematician

Objects as Points

2019-04-16
Xingyi Zhou , Dequan Wang , Philipp Krähenbühl
Detection identifies objects as axis-aligned boxes in an image. Most successful object detectors enumerate a nearly exhaustive list of potential object locations and classify each. This is wasteful, inefficient, and … Detection identifies objects as axis-aligned boxes in an image. Most successful object detectors enumerate a nearly exhaustive list of potential object locations and classify each. This is wasteful, inefficient, and requires additional post-processing. In this paper, we take a different approach. We model an object as a single point --- the center point of its bounding box. Our detector uses keypoint estimation to find center points and regresses to all other object properties, such as size, 3D location, orientation, and even pose. Our center point based approach, CenterNet, is end-to-end differentiable, simpler, faster, and more accurate than corresponding bounding box based detectors. CenterNet achieves the best speed-accuracy trade-off on the MS COCO dataset, with 28.1% AP at 142 FPS, 37.4% AP at 52 FPS, and 45.1% AP with multi-scale testing at 1.4 FPS. We use the same approach to estimate 3D bounding box in the KITTI benchmark and human pose on the COCO keypoint dataset. Our method performs competitively with sophisticated multi-stage methods and runs in real-time.

FCNs in the Wild: Pixel-level Adversarial and Constraint-based Adaptation

2016-01-01
Judy Hoffman , Dequan Wang , Fisher Yu +1 more
Fully convolutional models for dense prediction have proven successful for a wide range of visual tasks. Such models perform well in a supervised setting, but performance can be surprisingly poor … Fully convolutional models for dense prediction have proven successful for a wide range of visual tasks. Such models perform well in a supervised setting, but performance can be surprisingly poor under domain shifts that appear mild to a human observer. For example, training on one city and testing on another in a different geographic region and/or weather condition may result in significantly degraded performance due to pixel-level distribution shift. In this paper, we introduce the first domain adaptive semantic segmentation method, proposing an unsupervised adversarial approach to pixel prediction problems. Our method consists of both global and category specific adaptation techniques. Global domain alignment is performed using a novel semantic segmentation network with fully convolutional domain adversarial learning. This initially adapted space then enables category specific adaptation through a generalization of constrained weak learning, with explicit transfer of the spatial layout from the source to the target domains. Our approach outperforms baselines across different settings on multiple large-scale datasets, including adapting across various real city environments, different synthetic sub-domains, from simulated to real environments, and on a novel large-scale dash-cam dataset.

VisDA: The Visual Domain Adaptation Challenge

2017-01-01
Xingchao Peng , Ben Usman , Neela Kaushik +3 more
We present the 2017 Visual Domain Adaptation (VisDA) dataset and challenge, a large-scale testbed for unsupervised domain adaptation across visual domains. Unsupervised domain adaptation aims to solve the real-world problem … We present the 2017 Visual Domain Adaptation (VisDA) dataset and challenge, a large-scale testbed for unsupervised domain adaptation across visual domains. Unsupervised domain adaptation aims to solve the real-world problem of domain shift, where machine learning models trained on one domain must be transferred and adapted to a novel visual domain without additional supervision. The VisDA2017 challenge is focused on the simulation-to-reality shift and has two associated tasks: image classification and image segmentation. The goal in both tracks is to first train a model on simulated, synthetic data in the source domain and then adapt it to perform well on real image data in the unlabeled test domain. Our dataset is the largest one to date for cross-domain object classification, with over 280K images across 12 categories in the combined training, validation and testing domains. The image segmentation dataset is also large-scale with over 30K images across 18 categories in the three domains. We compare VisDA to existing cross-domain adaptation datasets and provide a baseline performance analysis using various domain adaptation models that are currently popular in the field.
View PDF Chat

Objects as Points

2019-01-01
Xingyi Zhou , Dequan Wang , Philipp Krähenbühl
Detection identifies objects as axis-aligned boxes in an image. Most successful object detectors enumerate a nearly exhaustive list of potential object locations and classify each. This is wasteful, inefficient, and … Detection identifies objects as axis-aligned boxes in an image. Most successful object detectors enumerate a nearly exhaustive list of potential object locations and classify each. This is wasteful, inefficient, and requires additional post-processing. In this paper, we take a different approach. We model an object as a single point --- the center point of its bounding box. Our detector uses keypoint estimation to find center points and regresses to all other object properties, such as size, 3D location, orientation, and even pose. Our center point based approach, CenterNet, is end-to-end differentiable, simpler, faster, and more accurate than corresponding bounding box based detectors. CenterNet achieves the best speed-accuracy trade-off on the MS COCO dataset, with 28.1% AP at 142 FPS, 37.4% AP at 52 FPS, and 45.1% AP with multi-scale testing at 1.4 FPS. We use the same approach to estimate 3D bounding box in the KITTI benchmark and human pose on the COCO keypoint dataset. Our method performs competitively with sophisticated multi-stage methods and runs in real-time.

Joint Monocular 3D Vehicle Detection and Tracking

2019-10-01
Hou-Ning Hu , Qi-Zhi Cai , Dequan Wang +5 more
Vehicle 3D extents and trajectories are critical cues for predicting the future location of vehicles and planning future agent ego-motion based on those predictions. In this paper, we propose a … Vehicle 3D extents and trajectories are critical cues for predicting the future location of vehicles and planning future agent ego-motion based on those predictions. In this paper, we propose a novel online framework for 3D vehicle detection and tracking from monocular videos. The framework can not only associate detections of vehicles in motion over time, but also estimate their complete 3D bounding box information from a sequence of 2D images captured on a moving platform. Our method leverages 3D box depth-ordering matching for robust instance association and utilizes 3D trajectory prediction for re-identification of occluded vehicles. We also design a motion learning module based on an LSTM for more accurate long-term motion extrapolation. Our experiments on simulation, KITTI, and Argoverse datasets show that our 3D tracking pipeline offers robust data association and tracking. On Argoverse, our image-based method is significantly better for tracking 3D vehicles within 30 meters than the LiDAR-centric baseline methods.
View PDF Chat

Tent: Fully Test-time Adaptation by Entropy Minimization

2020-01-01
Dequan Wang , Evan Shelhamer , Shaoteng Liu +2 more
A model must adapt itself to generalize to new and different data during testing. In this setting of fully test-time adaptation the model has only the test data and its … A model must adapt itself to generalize to new and different data during testing. In this setting of fully test-time adaptation the model has only the test data and its own parameters. We propose to adapt by test entropy minimization (tent): we optimize the model for confidence as measured by the entropy of its predictions. Our method estimates normalization statistics and optimizes channel-wise affine transformations to update online on each batch. Tent reduces generalization error for image classification on corrupted ImageNet and CIFAR-10/100 and reaches a new state-of-the-art error on ImageNet-C. Tent handles source-free domain adaptation on digit recognition from SVHN to MNIST/MNIST-M/USPS, on semantic segmentation from GTA to Cityscapes, and on the VisDA-C benchmark. These results are achieved in one epoch of test-time optimization without altering training.

Contrastive Test-Time Adaptation

2022-06-01
Dian Chen , Dequan Wang , Trevor Darrell +1 more
Test-time adaptation is a special setting of unsupervised domain adaptation where a trained model on the source domain has to adapt to the target domain without accessing source data. We … Test-time adaptation is a special setting of unsupervised domain adaptation where a trained model on the source domain has to adapt to the target domain without accessing source data. We propose a novel way to leverage self-supervised contrastive learning to facilitate target feature learning, along with an online pseudo labeling scheme with refinement that significantly denoises pseudo labels. The contrastive learning task is applied jointly with pseudo labeling, contrasting positive and negative pairs constructed similarly as MoCo but with source-initialized encoder, and excluding same-class negative pairs indicated by pseudo labels. Meanwhile, we produce pseudo labels online and refine them via soft voting among their nearest neighbors in the target feature space, enabled by maintaining a memory queue. Our method, AdaContrast, achieves state-of-the-art performance on major benchmarks while having several desirable properties compared to existing works, including memory efficiency, insensitivity to hyper-parameters, and better model calibration. Code is released at https://github.com/DianCh/AdaContrast.
View PDF Chat

Deep Object-Centric Policies for Autonomous Driving

2019-05-01
Dequan Wang , Coline Devin , Qi-Zhi Cai +2 more
While learning visuomotor skills in an end-to-end manner is appealing, deep neural networks are often uninterpretable and fail in surprising ways. For robotics tasks, such as autonomous driving, models that … While learning visuomotor skills in an end-to-end manner is appealing, deep neural networks are often uninterpretable and fail in surprising ways. For robotics tasks, such as autonomous driving, models that explicitly represent objects may be more robust to new scenes and provide intuitive visualizations. We describe a taxonomy of “object-centric” models which leverage both object instances and end-to-end learning. In the Grand Theft Auto V simulator, we show that object-centric models outperform object-agnostic methods in scenes with other vehicles and pedestrians, even with an imperfect detector. We also demonstrate that our architectures perform well on real-world environments by evaluating on the Berkeley DeepDrive Video dataset, where an object-centric model outperforms object-agnostic models in the low-data regimes.
View PDF Chat

Monocular Plan View Networks for Autonomous Driving

2019-11-01
Dequan Wang , Coline Devin , Qi-Zhi Cai +2 more
Convolutions on monocular dash cam videos capture spatial invariances in the image plane but do not explicitly reason about distances and depth. We propose a simple transformation of observations into … Convolutions on monocular dash cam videos capture spatial invariances in the image plane but do not explicitly reason about distances and depth. We propose a simple transformation of observations into a bird's eye view, also known as plan view, for end-to-end control. We detect vehicles and pedestrians in the first person view and project them into an overhead plan view. This representation provides an abstraction of the environment from which a deep network can easily deduce the positions and directions of entities. Additionally, the plan view enables us to leverage advances in 3D object detection in conjunction with deep policy learning. We evaluate our monocular plan view network on the photo-realistic Grand Theft Auto V simulator. A network using both a plan view and front view causes less than half as many collisions as previous detection-based methods and an order of magnitude fewer collisions than pure pixel-based policies.
View PDF Chat

Tent: Fully Test-time Adaptation by Entropy Minimization

2020-06-18
Dequan Wang , Evan Shelhamer , Shaoteng Liu +2 more
A model must adapt itself to generalize to new and different data during testing. In this setting of fully test-time adaptation the model has only the test data and its … A model must adapt itself to generalize to new and different data during testing. In this setting of fully test-time adaptation the model has only the test data and its own parameters. We propose to adapt by test entropy minimization (tent): we optimize the model for confidence as measured by the entropy of its predictions. Our method estimates normalization statistics and optimizes channel-wise affine transformations to update online on each batch. Tent reduces generalization error for image classification on corrupted ImageNet and CIFAR-10/100 and reaches a new state-of-the-art error on ImageNet-C. Tent handles source-free domain adaptation on digit recognition from SVHN to MNIST/MNIST-M/USPS, on semantic segmentation from GTA to Cityscapes, and on the VisDA-C benchmark. These results are achieved in one epoch of test-time optimization without altering training.
View PDF Chat

CoDeNet: Efficient Deployment of Input-Adaptive Object Detection on Embedded FPGAs

2021-02-17
Qijing Huang , Dequan Wang , Zhen Dong +6 more
Deploying deep learning models on embedded systems for computer vision tasks has been challenging due to limited compute resources and strict energy budgets. The majority of existing work focuses on … Deploying deep learning models on embedded systems for computer vision tasks has been challenging due to limited compute resources and strict energy budgets. The majority of existing work focuses on accelerating image classification, while other fundamental vision problems, such as object detection, have not been adequately addressed. Compared with image classification, detection problems are more sensitive to the spatial variance of objects, and therefore, require specialized convolutions to aggregate spatial information. To address this need, recent work introduces dynamic deformable convolution to augment regular convolutions. Regular convolutions process a fixed grid of pixels across all the spatial locations in an image, while dynamic deformable convolution may access arbitrary pixels in the image with the access pattern being input-dependent and varying with spatial location. These properties lead to inefficient memory accesses of inputs with existing hardware.
View PDF Chat

Joint Monocular 3D Vehicle Detection and Tracking

2018-01-01
Hou-Ning Hu , Qi-Zhi Cai , Dequan Wang +5 more
Vehicle 3D extents and trajectories are critical cues for predicting the future location of vehicles and planning future agent ego-motion based on those predictions. In this paper, we propose a … Vehicle 3D extents and trajectories are critical cues for predicting the future location of vehicles and planning future agent ego-motion based on those predictions. In this paper, we propose a novel online framework for 3D vehicle detection and tracking from monocular videos. The framework can not only associate detections of vehicles in motion over time, but also estimate their complete 3D bounding box information from a sequence of 2D images captured on a moving platform. Our method leverages 3D box depth-ordering matching for robust instance association and utilizes 3D trajectory prediction for re-identification of occluded vehicles. We also design a motion learning module based on an LSTM for more accurate long-term motion extrapolation. Our experiments on simulation, KITTI, and Argoverse datasets show that our 3D tracking pipeline offers robust data association and tracking. On Argoverse, our image-based method is significantly better for tracking 3D vehicles within 30 meters than the LiDAR-centric baseline methods.
View PDF Chat

BEV-Seg: Bird's Eye View Semantic Segmentation Using Geometry and Semantic Point Cloud

2020-01-01
Mong H. Ng , Kaahan Radia , Chen Jian-fei +3 more
Bird's-eye-view (BEV) is a powerful and widely adopted representation for road scenes that captures surrounding objects and their spatial locations, along with overall context in the scene. In this work, … Bird's-eye-view (BEV) is a powerful and widely adopted representation for road scenes that captures surrounding objects and their spatial locations, along with overall context in the scene. In this work, we focus on bird's eye semantic segmentation, a task that predicts pixel-wise semantic segmentation in BEV from side RGB images. This task is made possible by simulators such as Carla, which allow for cheap data collection, arbitrary camera placements, and supervision in ways otherwise not possible in the real world. There are two main challenges to this task: the view transformation from side view to bird's eye view, as well as transfer learning to unseen domains. Existing work transforms between views through fully connected layers and transfer learns via GANs. This suffers from a lack of depth reasoning and performance degradation across domains. Our novel 2-staged perception pipeline explicitly predicts pixel depths and combines them with pixel semantics in an efficient manner, allowing the model to leverage depth information to infer objects' spatial locations in the BEV. In addition, we transfer learning by abstracting high-level geometric features and predicting an intermediate representation that is common across different domains. We publish a new dataset called BEVSEG-Carla and show that our approach improves state-of-the-art by 24% mIoU and performs well when transferred to a new domain.

Deep Layer Aggregation

2017-01-01
Fisher Yu , Dequan Wang , Evan Shelhamer +1 more
Visual recognition requires rich representations that span levels from low to high, scales from small to large, and resolutions from fine to coarse. Even with the depth of features in … Visual recognition requires rich representations that span levels from low to high, scales from small to large, and resolutions from fine to coarse. Even with the depth of features in a convolutional network, a layer in isolation is not enough: compounding and aggregating these representations improves inference of what and where. Architectural efforts are exploring many dimensions for network backbones, designing deeper or wider architectures, but how to best aggregate layers and blocks across a network deserves further attention. Although skip connections have been incorporated to combine layers, these connections have been "shallow" themselves, and only fuse by simple, one-step operations. We augment standard architectures with deeper aggregation to better fuse information across layers. Our deep layer aggregation structures iteratively and hierarchically merge the feature hierarchy to make networks with better accuracy and fewer parameters. Experiments across architectures and tasks show that deep layer aggregation improves recognition and resolution compared to existing branching and merging schemes. The code is at https://github.com/ucbdrive/dla.

Fully Test-time Adaptation by Entropy Minimization

2020-06-18
Dequan Wang , Evan Shelhamer , Shaoteng Liu +2 more
A model must adapt itself to generalize to new and different data during testing. In this setting of fully test-time adaptation the model has only the test data and its … A model must adapt itself to generalize to new and different data during testing. In this setting of fully test-time adaptation the model has only the test data and its own parameters. We propose to adapt by test entropy minimization (tent): we optimize the model for confidence as measured by the entropy of its predictions. Our method estimates normalization statistics and optimizes channel-wise affine transformations to update online on each batch. Tent reduces generalization error for image classification on corrupted ImageNet and CIFAR-10/100 and reaches a new state-of-the-art error on ImageNet-C. Tent handles source-free domain adaptation on digit recognition from SVHN to MNIST/MNIST-M/USPS, on semantic segmentation from GTA to Cityscapes, and on the VisDA-C benchmark. These results are achieved in one epoch of test-time optimization without altering training.

Blurring the Line Between Structure and Learning to Optimize and Adapt Receptive Fields.

2019-04-25
Evan Shelhamer , Dequan Wang , Trevor Darrell
The visual world is vast and varied, but its variations divide into structured and unstructured factors. We compose free-form filters and structured Gaussian filters, optimized end-to-end, to factorize deep representations … The visual world is vast and varied, but its variations divide into structured and unstructured factors. We compose free-form filters and structured Gaussian filters, optimized end-to-end, to factorize deep representations and learn both local features and their degree of locality. Our semi-structured composition is strictly more expressive than free-form filtering, and changes in its structured parameters would require changes in free-form architecture. In effect this optimizes over receptive field size and shape, tuning locality to the data and task. Dynamic inference, in which the Gaussian structure varies with the input, adapts receptive field size to compensate for local scale variation. Optimizing receptive field size improves semantic segmentation accuracy on Cityscapes by 1-2 points for strong dilated and skip architectures and by up to 10 points for suboptimal designs. Adapting receptive fields by dynamic Gaussian structure further improves results, equaling the accuracy of free-form deformation while improving efficiency.

Back to the Source: Diffusion-Driven Test-Time Adaptation

2022-01-01
Jin Gao , Jialing Zhang , Xihui Liu +3 more
Test-time adaptation harnesses test inputs to improve the accuracy of a model trained on source data when tested on shifted target data. Existing methods update the source model by (re-)training … Test-time adaptation harnesses test inputs to improve the accuracy of a model trained on source data when tested on shifted target data. Existing methods update the source model by (re-)training on each target domain. While effective, re-training is sensitive to the amount and order of the data and the hyperparameters for optimization. We instead update the target data, by projecting all test inputs toward the source domain with a generative diffusion model. Our diffusion-driven adaptation method, DDA, shares its models for classification and generation across all domains. Both models are trained on the source domain, then fixed during testing. We augment diffusion with image guidance and self-ensembling to automatically decide how much to adapt. Input adaptation by DDA is more robust than prior model adaptation approaches across a variety of corruptions, architectures, and data regimes on the ImageNet-C benchmark. With its input-wise updates, DDA succeeds where model adaptation degrades on too little data in small batches, dependent data in non-uniform order, or mixed data with multiple corruptions.
View PDF Chat

Blurring the Line Between Structure and Learning to Optimize and Adapt Receptive Fields

2019-01-01
Evan Shelhamer , Dequan Wang , Trevor Darrell
The visual world is vast and varied, but its variations divide into structured and unstructured factors. We compose free-form filters and structured Gaussian filters, optimized end-to-end, to factorize deep representations … The visual world is vast and varied, but its variations divide into structured and unstructured factors. We compose free-form filters and structured Gaussian filters, optimized end-to-end, to factorize deep representations and learn both local features and their degree of locality. Our semi-structured composition is strictly more expressive than free-form filtering, and changes in its structured parameters would require changes in free-form architecture. In effect this optimizes over receptive field size and shape, tuning locality to the data and task. Dynamic inference, in which the Gaussian structure varies with the input, adapts receptive field size to compensate for local scale variation. Optimizing receptive field size improves semantic segmentation accuracy on Cityscapes by 1-2 points for strong dilated and skip architectures and by up to 10 points for suboptimal designs. Adapting receptive fields by dynamic Gaussian structure further improves results, equaling the accuracy of free-form deformation while improving efficiency.
View PDF Chat

Data-Centric Foundation Models in Computational Healthcare: A Survey

2024-01-01
Yunkun Zhang , Jin Gao , Zheling Tan +5 more
The advent of foundation models (FMs) as an emerging suite of AI techniques has struck a wave of opportunities in computational healthcare. The interactive nature of these models, guided by … The advent of foundation models (FMs) as an emerging suite of AI techniques has struck a wave of opportunities in computational healthcare. The interactive nature of these models, guided by pre-training data and human instructions, has ignited a data-centric AI paradigm that emphasizes better data characterization, quality, and scale. In healthcare AI, obtaining and processing high-quality clinical data records has been a longstanding challenge, ranging from data quantity, annotation, patient privacy, and ethics. In this survey, we investigate a wide range of data-centric approaches in the FM era (from model pre-training to inference) towards improving the healthcare workflow. We discuss key perspectives in AI security, assessment, and alignment with human values. Finally, we offer a promising outlook of FM-based analytics to enhance the performance of patient outcome and clinical workflow in the evolving landscape of healthcare and medicine. We provide an up-to-date list of healthcare-related foundation models and datasets at https://github.com/Yunkun-Zhang/Data-Centric-FM-Healthcare .

Deep Object-Centric Policies for Autonomous Driving

2018-01-01
Dequan Wang , Coline Devin , Qi-Zhi Cai +2 more
While learning visuomotor skills in an end-to-end manner is appealing, deep neural networks are often uninterpretable and fail in surprising ways. For robotics tasks, such as autonomous driving, models that … While learning visuomotor skills in an end-to-end manner is appealing, deep neural networks are often uninterpretable and fail in surprising ways. For robotics tasks, such as autonomous driving, models that explicitly represent objects may be more robust to new scenes and provide intuitive visualizations. We describe a taxonomy of "object-centric" models which leverage both object instances and end-to-end learning. In the Grand Theft Auto V simulator, we show that object-centric models outperform object-agnostic methods in scenes with other vehicles and pedestrians, even with an imperfect detector. We also demonstrate that our architectures perform well on real-world environments by evaluating on the Berkeley DeepDrive Video dataset, where an object-centric model outperforms object-agnostic models in the low-data regimes.
View PDF Chat

Monocular Plan View Networks for Autonomous Driving

2019-05-16
Dequan Wang , Coline Devin , Qi-Zhi Cai +2 more
Convolutions on monocular dash cam videos capture spatial invariances in the image plane but do not explicitly reason about distances and depth. We propose a simple transformation of observations into … Convolutions on monocular dash cam videos capture spatial invariances in the image plane but do not explicitly reason about distances and depth. We propose a simple transformation of observations into a bird's eye view, also known as plan view, for end-to-end control. We detect vehicles and pedestrians in the first person view and project them into an overhead plan view. This representation provides an abstraction of the environment from which a deep network can easily deduce the positions and directions of entities. Additionally, the plan view enables us to leverage advances in 3D object detection in conjunction with deep policy learning. We evaluate our monocular plan view network on the photo-realistic Grand Theft Auto V simulator. A network using both a plan view and front view causes less than half as many collisions as previous detection-based methods and an order of magnitude fewer collisions than pure pixel-based policies.

Fighting Gradients with Gradients: Dynamic Defenses against Adversarial Attacks

2021-01-01
Dequan Wang , An Ju , Evan Shelhamer +2 more
Adversarial attacks optimize against models to defeat defenses. Existing defenses are static, and stay the same once trained, even while attacks change. We argue that models should fight back, and … Adversarial attacks optimize against models to defeat defenses. Existing defenses are static, and stay the same once trained, even while attacks change. We argue that models should fight back, and optimize their defenses against attacks at test time. We propose dynamic defenses, to adapt the model and input during testing, by defensive entropy minimization (dent). Dent alters testing, but not training, for compatibility with existing models and train-time defenses. Dent improves the robustness of adversarially-trained defenses and nominally-trained models against white-box, black-box, and adaptive attacks on CIFAR-10/100 and ImageNet. In particular, dent boosts state-of-the-art defenses by 20+ points absolute against AutoAttack on CIFAR-10 at $\epsilon_\infty$ = 8/255.
View PDF Chat

Convolutional Neural Networks on non-uniform geometrical signals using Euclidean spectral transformation

2019-01-01
Chiyu Jiang , Dequan Wang , Jingwei Huang +2 more
Convolutional Neural Networks (CNN) have been successful in processing data signals that are uniformly sampled in the spatial domain (e.g., images). However, most data signals do not natively exist on … Convolutional Neural Networks (CNN) have been successful in processing data signals that are uniformly sampled in the spatial domain (e.g., images). However, most data signals do not natively exist on a grid, and in the process of being sampled onto a uniform physical grid suffer significant aliasing error and information loss. Moreover, signals can exist in different topological structures as, for example, points, lines, surfaces and volumes. It has been challenging to analyze signals with mixed topologies (for example, point cloud with surface mesh). To this end, we develop mathematical formulations for Non-Uniform Fourier Transforms (NUFT) to directly, and optimally, sample nonuniform data signals of different topologies defined on a simplex mesh into the spectral domain with no spatial sampling error. The spectral transform is performed in the Euclidean space, which removes the translation ambiguity from works on the graph spectrum. Our representation has four distinct advantages: (1) the process causes no spatial sampling error during the initial sampling, (2) the generality of this approach provides a unified framework for using CNNs to analyze signals of mixed topologies, (3) it allows us to leverage state-of-the-art backbone CNN architectures for effective learning without having to design a particular architecture for a particular data structure in an ad-hoc fashion, and (4) the representation allows weighted meshes where each element has a different weight (i.e., texture) indicating local properties. We achieve results on par with the state-of-the-art for the 3D shape retrieval task, and a new state-of-the-art for the point cloud to surface reconstruction task.
View PDF Chat

Towards General Purpose Medical AI: Continual Learning Medical Foundation Model

2023-01-01
Huahui Yi , Ziyuan Qin , Qicheng Lao +5 more
Inevitable domain and task discrepancies in real-world scenarios can impair the generalization performance of the pre-trained deep models for medical data. Therefore, we audaciously propose that we should build a … Inevitable domain and task discrepancies in real-world scenarios can impair the generalization performance of the pre-trained deep models for medical data. Therefore, we audaciously propose that we should build a general-purpose medical AI system that can be seamlessly adapted to downstream domains/tasks. Since the domain/task adaption procedures usually involve additional labeling work for the target data, designing a data-efficient adaption algorithm is desired to save the cost of transferring the learned knowledge. Our recent work found that vision-language models (VLMs) are efficient learners with extraordinary cross-domain ability. Therefore, in this work, we further explore the possibility of leveraging pre-trained VLMs as medical foundation models for building general-purpose medical AI, where we thoroughly investigate three machine-learning paradigms, i.e., domain/task-specialized learning, joint learning, and continual learning, for training the VLMs and evaluate their generalization performance on cross-domain and cross-task test sets. To alleviate the catastrophic forgetting during sequential training, we employ rehearsal learning and receive a sharp boost in terms of generalization capability. In a nutshell, our empirical evidence suggests that continual learning may be a practical and efficient learning paradigm for the medical foundation model. And we hope researchers can use our empirical evidence as basement to further explore the path toward medical foundation model.

Dynamic Scale Inference by Entropy Minimization

2019-01-01
Dequan Wang , Evan Shelhamer , Bruno A. Olshausen +1 more
Given the variety of the visual world there is not one true scale for recognition: objects may appear at drastically different sizes across the visual field. Rather than enumerate variations … Given the variety of the visual world there is not one true scale for recognition: objects may appear at drastically different sizes across the visual field. Rather than enumerate variations across filter channels or pyramid levels, dynamic models locally predict scale and adapt receptive fields accordingly. The degree of variation and diversity of inputs makes this a difficult task. Existing methods either learn a feedforward predictor, which is not itself totally immune to the scale variation it is meant to counter, or select scales by a fixed algorithm, which cannot learn from the given task and data. We extend dynamic scale inference from feedforward prediction to iterative optimization for further adaptivity. We propose a novel entropy minimization objective for inference and optimize over task and structure parameters to tune the model to each input. Optimization during inference improves semantic segmentation accuracy and generalizes better to extreme scale variations that cause feedforward dynamic inference to falter.

Algorithm-hardware Co-design for Deformable Convolution

2019-12-01
Qijing Huang , Dequan Wang , Yizhao Gao +5 more
FPGAs provide a flexible and efficient platform to accelerate rapidly-changing algorithms for computer vision. The majority of existing work focuses on accelerating image classification, while other fundamental vision problems, including … FPGAs provide a flexible and efficient platform to accelerate rapidly-changing algorithms for computer vision. The majority of existing work focuses on accelerating image classification, while other fundamental vision problems, including object detection and instance segmentation, have not been adequately addressed. Compared with image classification, detection problems are more sensitive to the spatial variance of objects, and therefore, require specialized convolutions to aggregate spatial information. To address this, recent work proposes dynamic deformable convolution to augment regular convolutions. Regular convolutions process a fixed grid of pixels across all the spatial locations in an image, while dynamic deformable convolutions may access arbitrary pixels. The access pattern of deformable convolutions is input-dependent and varies per spatial location. These properties lead to inefficient memory accesses of inputs with existing hardware. In this work, we first investigate the overhead of the deformable convolution on embedded FPGA SoCs, and then show the accuracy-latency tradeoffs for a set of algorithm modifications, including full versus depthwise, fixed-shape, and limited-range. These modifications benefit the efficiency of the embedded accelerator in general. We build an efficient object detection network with modified deformable convolutions and quantize the network using state-of-the-art quantization methods. Experiments show that our co-design optimization for the deformable convolution achieves significant hardware speedup with little accuracy compromised.
View PDF Chat

Convolutional Neural Networks on non-uniform geometrical signals using Euclidean spectral transformation

2019-01-07
Chiyu Max Jiang , Dequan Wang , Jingwei Huang +2 more
Convolutional Neural Networks (CNN) have been successful in processing data signals that are uniformly sampled in the spatial domain (e.g., images). However, most data signals do not natively exist on … Convolutional Neural Networks (CNN) have been successful in processing data signals that are uniformly sampled in the spatial domain (e.g., images). However, most data signals do not natively exist on a grid, and in the process of being sampled onto a uniform physical grid suffer significant aliasing error and information loss. Moreover, signals can exist in different topological structures as, for example, points, lines, surfaces and volumes. It has been challenging to analyze signals with mixed topologies (for example, point cloud with surface mesh). To this end, we develop mathematical formulations for Non-Uniform Fourier Transforms (NUFT) to directly, and optimally, sample nonuniform data signals of different topologies defined on a simplex mesh into the spectral domain with no spatial sampling error. The spectral transform is performed in the Euclidean space, which removes the translation ambiguity from works on the graph spectrum. Our representation has four distinct advantages: (1) the process causes no spatial sampling error during the initial sampling, (2) the generality of this approach provides a unified framework for using CNNs to analyze signals of mixed topologies, (3) it allows us to leverage state-of-the-art backbone CNN architectures for effective learning without having to design a particular architecture for a particular data structure in an ad-hoc fashion, and (4) the representation allows weighted meshes where each element has a different weight (i.e., texture) indicating local properties. We achieve results on par with the state-of-the-art for the 3D shape retrieval task, and a new state-of-the-art for the point cloud to surface reconstruction task.

Modified Homotopy Analysis Method for Nonlinear Aeroelastic Behavior of Two Degree-of-Freedom Airfoils

2015-10-07
Yaobin Niu , Zhongwei Wang , Dequan Wang +1 more
In this paper, the homotopy analysis method (HAM) is extended to deal with the nonlinear aeroelastic problem of a two degree-of-freedom (DOF) airfoil. To avoid determination of the parameter for … In this paper, the homotopy analysis method (HAM) is extended to deal with the nonlinear aeroelastic problem of a two degree-of-freedom (DOF) airfoil. To avoid determination of the parameter for the complicated high-order minimization problem, a new modified HAM is proposed based on the idea of minimizing the squared residual. Using this method, the convergence-control parameter is determined by the low order squared residual of the governing equations, and then the problem is solved in a way similar to the basic HAM. The proposed method is used to solve the nonlinear aeroelastic behavior of a supersonic airfoil, with the unsteady aerodynamic load evaluated by the piston theory. Two examples are prepared, for which the frequencies and amplitudes of the limit cycles are obtained. The approximate solutions obtained are demonstrated to agree excellently the numerical solutions, meanwhile, the convergence-control parameter can be easily determined using the present approach.

Learning to Point and Count

2015-01-01
Jie Shao , Dequan Wang , Xiangyang Xue +1 more
This paper proposes the problem of point-and-count as a test case to break the what-and-where deadlock. Different from the traditional detection problem, the goal is to discover key salient points … This paper proposes the problem of point-and-count as a test case to break the what-and-where deadlock. Different from the traditional detection problem, the goal is to discover key salient points as a way to localize and count the number of objects simultaneously. We propose two alternatives, one that counts first and then point, and another that works the other way around. Fundamentally, they pivot around whether we solve "what" or "where" first. We evaluate their performance on dataset that contains multiple instances of the same class, demonstrating the potentials and their synergies. The experiences derive a few important insights that explains why this is a much harder problem than classification, including strong data bias and the inability to deal with object scales robustly in state-of-art convolutional neural networks.

Convolutional Neural Networks on Non-uniform Geometrical Signals Using Euclidean Spectral Transformation

2019-01-01
Chiyu Max Jiang , Dequan Wang , Jingwei Huang +2 more
Convolutional Neural Networks (CNN) have been successful in processing data signals that are uniformly sampled in the spatial domain (e.g., images). However, most data signals do not natively exist on … Convolutional Neural Networks (CNN) have been successful in processing data signals that are uniformly sampled in the spatial domain (e.g., images). However, most data signals do not natively exist on a grid, and in the process of being sampled onto a uniform physical grid suffer significant aliasing error and information loss. Moreover, signals can exist in different topological structures as, for example, points, lines, surfaces and volumes. It has been challenging to analyze signals with mixed topologies (for example, point cloud with surface mesh). To this end, we develop mathematical formulations for Non-Uniform Fourier Transforms (NUFT) to directly, and optimally, sample nonuniform data signals of different topologies defined on a simplex mesh into the spectral domain with no spatial sampling error. The spectral transform is performed in the Euclidean space, which removes the translation ambiguity from works on the graph spectrum. Our representation has four distinct advantages: (1) the process causes no spatial sampling error during the initial sampling, (2) the generality of this approach provides a unified framework for using CNNs to analyze signals of mixed topologies, (3) it allows us to leverage state-of-the-art backbone CNN architectures for effective learning without having to design a particular architecture for a particular data structure in an ad-hoc fashion, and (4) the representation allows weighted meshes where each element has a different weight (i.e., texture) indicating local properties. We achieve results on par with the state-of-the-art for the 3D shape retrieval task, and a new state-of-the-art for the point cloud to surface reconstruction task.

Algorithm-hardware Co-design for Deformable Convolution

2020-01-01
Qijing Huang , Dequan Wang , Yizhao Gao +5 more
FPGAs provide a flexible and efficient platform to accelerate rapidly-changing algorithms for computer vision. The majority of existing work focuses on accelerating image classification, while other fundamental vision problems, including … FPGAs provide a flexible and efficient platform to accelerate rapidly-changing algorithms for computer vision. The majority of existing work focuses on accelerating image classification, while other fundamental vision problems, including object detection and instance segmentation, have not been adequately addressed. Compared with image classification, detection problems are more sensitive to the spatial variance of objects, and therefore, require specialized convolutions to aggregate spatial information. To address this, recent work proposes dynamic deformable convolution to augment regular convolutions. Regular convolutions process a fixed grid of pixels across all the spatial locations in an image, while dynamic deformable convolutions may access arbitrary pixels in the image and the access pattern is input-dependent and varies per spatial location. These properties lead to inefficient memory accesses of inputs with existing hardware. In this work, we first investigate the overhead of the deformable convolution on embedded FPGA SoCs, and then show the accuracy-latency tradeoffs for a set of algorithm modifications including full versus depthwise, fixed-shape, and limited-range. These modifications benefit the energy efficiency for embedded devices in general as they reduce the compute complexity. We then build an efficient object detection network with modified deformable convolutions and quantize the network using state-of-the-art quantization methods. We implement a unified hardware engine on FPGA to support all the operations in the network. Preliminary experiments show that little accuracy is compromised and speedup can be achieved with our co-design optimization for the deformable convolution.
View PDF Chat

MAROAM: Map-based Radar SLAM through Two-step Feature Selection

2022-01-01
Dequan Wang , Yifan Duan , Xiaoran Fan +3 more
In this letter, we propose MAROAM, a millimeter wave radar-based SLAM framework, which employs a two-step feature selection process to build the global consistent map. Specifically, we first extract feature … In this letter, we propose MAROAM, a millimeter wave radar-based SLAM framework, which employs a two-step feature selection process to build the global consistent map. Specifically, we first extract feature points from raw data based on their local geometric properties to filter out those points that violate the principle of millimeter-wave radar imaging. Then, we further employ another round of probabilistic feature selection by examining how often and how recent the feature point has been detected in the proceeding frames. With such a two-step feature selection, we establish a global consistent map for accurate and robust pose estimation as well as other downstream tasks. At last, we perform loop closure and graph optimization in the back-end, further reducing the accumulated drift error. We evaluate the performance of MAROAM on the three datasets: the Oxford Radar RobotCar Dataset, the MulRan Dataset and the Boreas Dataset. We consider a variety of experimental settings with different scenery, weather, and road conditions. The experimental results show that the accuracy of MAROAM is 7.95%, 37.0% and 8.9% higher than the currently best-performing algorithms on these three datasets, respectively. The ablation results also show that our map-based odometry performs 28.6% better than the commonly used scan-to-frames method. Finally, as devoted contributors to the open-source community, we will open source the algorithm after the paper is accepted.
View PDF Chat

mmPlace: Robust Place Recognition With Intermediate Frequency Signal of Low-Cost Single-Chip Millimeter Wave Radar

2024-03-14
Chengzhen Meng , Yifan Duan , Chenming He +3 more
Place recognition is crucial for tasks like loop-closure detection and re-localization. Single-chip millimeter wave radar (single-chip radar in short) emerges as a low-cost sensor option for place recognition, with the … Place recognition is crucial for tasks like loop-closure detection and re-localization. Single-chip millimeter wave radar (single-chip radar in short) emerges as a low-cost sensor option for place recognition, with the advantage of insensitivity to degraded visual environments. However, it encounters two challenges. Firstly, sparse point cloud from single-chip radar leads to poor performance when using current place recognition methods, which assume much denser data. Secondly, its performance significantly declines in scenarios involving rotational and lateral variations, due to limited overlap in its field of view (FOV). We propose mmPlace, a robust place recognition system to address these challenges. Specifically, mmPlace transforms intermediate frequency (IF) signal into range azimuth heatmap and employs a spatial encoder to extract features. Additionally, to improve the performance in scenarios involving rotational and lateral variations, mmPlace employs a rotating platform and concatenates heatmaps in a rotation cycle, effectively expanding the system's FOV. We evaluate mmPlace's performance on the milliSonic dataset, which is collected on the University of Science and Technology of China (USTC) campus, the city roads surrounding the campus, and an underground parking garage. The results demonstrate that mmPlace outperforms point cloud-based methods and achieves 87.37% recall@1 in scenarios involving rotational and lateral variations.
View PDF Chat

Numerical Simulation of Fuel Regression Rate Behavior of Hybrid Rocket Motors

2013-01-01
Jianxin Hu , Zhi-xun Xia , Dequan Wang +1 more
The present study attempts to predict the performance of hybrid rocket motors using a physics-based comprehensive model in which all the submodels that govern the various processes that occur during … The present study attempts to predict the performance of hybrid rocket motors using a physics-based comprehensive model in which all the submodels that govern the various processes that occur during heating and combustion of solid fuel are based on physical-based mathematical models. Based on the fluid-solid coupling technique and some comprehensive physical processes during operation of hybrid rocket motors, a numerical model is developed to predict the regression rate for the solid fuel surface of the hybrid rocket motor under different operation conditions. The muti-dimensional Favre-averaged compressible turbulent Navier-Stokes equations are used as the governing equations of the reacting flow, the two-equation turbulence model is used to simulate the turbulent flow, and the eddy breakup model is used to simulate the gas combustion. The results presented are for hydroxyl-terminated-polybutadiene fuel and gaseous oxygen. The Navier-Stokes model results allow more detailed and realistic observation of the chamber flow field than is permitted by simpler boundary layer analyses. The model predications indicate that fuel surface regression rates are considerably impacted by both the size and geometry of the configuration. This study provides considerable information to the understanding of flow and combustion process in hybrid rocket motors. Methods for optimizing the hybrid combustion characteristic are put forward.
View PDF Chat

Iterative object and part transfer for fine-grained recognition

2017-07-01
Zhiqiang Shen , Yu–Gang Jiang , Dequan Wang +1 more
The aim of fine-grained recognition is to identify sub-ordinate categories in images like different species of birds. Existing works have confirmed that, in order to capture the subtle differences across … The aim of fine-grained recognition is to identify sub-ordinate categories in images like different species of birds. Existing works have confirmed that, in order to capture the subtle differences across the categories, automatic localization of objects and parts is critical. Most approaches for object and part localization rehed on the bottom-up pipeline, where thousands of region proposals are generated and then filtered by pre-trained object/part models. This is computationally expensive and not scalable once the number of objects/parts becomes large. In this paper, we propose a nonparametric data-driven method for object and part localization. Given an unlabeled test image, our approach transfers annotations from a few similar images retrieved in the training set. In particular, we propose an iterative transfer strategy that gradually refine the predicted bounding boxes. Based on the located objects and parts, deep convolutional features are extracted for recognition. We evaluate our approach on the widely-used CUB200-2011 dataset and a new and large dataset called Birdsnap. On both datasets, we achieve better results than many state-of-the-art approaches, including a few using oracle (manually annotated) bounding boxes in the test images.
View PDF Chat

GACT: Activation Compressed Training for Generic Network Architectures

2022-01-01
Xiaoxuan Liu , Lianmin Zheng , Dequan Wang +7 more
Training large neural network (NN) models requires extensive memory resources, and Activation Compressed Training (ACT) is a promising approach to reduce training memory footprint. This paper presents GACT, an ACT … Training large neural network (NN) models requires extensive memory resources, and Activation Compressed Training (ACT) is a promising approach to reduce training memory footprint. This paper presents GACT, an ACT framework to support a broad range of machine learning tasks for generic NN architectures with limited domain knowledge. By analyzing a linearized version of ACT's approximate gradient, we prove the convergence of GACT without prior knowledge on operator type or model architecture. To make training stable, we propose an algorithm that decides the compression ratio for each tensor by estimating its impact on the gradient at run time. We implement GACT as a PyTorch library that readily applies to any NN architecture. GACT reduces the activation memory for convolutional NNs, transformers, and graph NNs by up to 8.1x, enabling training with a 4.2x to 24.7x larger batch size, with negligible accuracy loss. We implement GACT as a PyTorch library at https://github.com/LiuXiaoxuanPKU/GACT-ICML.
View PDF Chat

CoDeNet: Algorithm-hardware Co-design for Deformable Convolution

2020-06-12
Zhen Dong , Dequan Wang , Qijing Huang +5 more
Deploying deep learning models on embedded systems for computer vision tasks has been challenging due to limited compute resources and strict energy budgets. The majority of existing work focuses on … Deploying deep learning models on embedded systems for computer vision tasks has been challenging due to limited compute resources and strict energy budgets. The majority of existing work focuses on accelerating image classification, while other fundamental vision problems, such as object detection, have not been adequately addressed. Compared with image classification, detection problems are more sensitive to the spatial variance of objects, and therefore, require specialized convolutions to aggregate spatial information. To address this, recent work proposes dynamic deformable convolution to augment regular convolutions. Regular convolutions process a fixed grid of pixels across all the spatial locations in an image, while dynamic deformable convolution may access arbitrary pixels in the image and the access pattern is input-dependent and varies per spatial location. These properties lead to inefficient memory accesses of inputs with existing hardware. In this work, we first investigate the overhead of the deformable convolution on embedded FPGA SoCs, and introduce a depthwise deformable convolution to reduce the total number of operations required. We then show the speed-accuracy tradeoffs for a set of algorithm modifications including irregular-access versus limited-range and fixed-shape. We evaluate these algorithmic changes with corresponding hardware optimizations. Results show a 1.36x and 9.76x speedup respectively for the full and depthwise deformable convolution on the embedded FPGA accelerator with minor accuracy loss on the object detection task. We then co-design an efficient network CoDeNet with the modified deformable convolution for object detection and quantize the network to 4-bit weights and 8-bit activations. Results show that our designs lie on the pareto-optimal front of the latency-accuracy tradeoff for the object detection task on embedded FPGAs

Monocular Plan View Networks for Autonomous Driving

2019-01-01
Dequan Wang , Devin Coline , Qi-Zhi Cai +2 more
Convolutions on monocular dash cam videos capture spatial invariances in the image plane but do not explicitly reason about distances and depth. We propose a simple transformation of observations into … Convolutions on monocular dash cam videos capture spatial invariances in the image plane but do not explicitly reason about distances and depth. We propose a simple transformation of observations into a bird's eye view, also known as plan view, for end-to-end control. We detect vehicles and pedestrians in the first person view and project them into an overhead plan view. This representation provides an abstraction of the environment from which a deep network can easily deduce the positions and directions of entities. Additionally, the plan view enables us to leverage advances in 3D object detection in conjunction with deep policy learning. We evaluate our monocular plan view network on the photo-realistic Grand Theft Auto V simulator. A network using both a plan view and front view causes less than half as many collisions as previous detection-based methods and an order of magnitude fewer collisions than pure pixel-based policies.
View PDF Chat

ActNN: Reducing Training Memory Footprint via 2-Bit Activation Compressed Training

2021-01-01
Jianfei Chen , Lianmin Zheng , Zhewei Yao +4 more
The increasing size of neural network models has been critical for improvements in their accuracy, but device memory is not growing at the same rate. This creates fundamental challenges for … The increasing size of neural network models has been critical for improvements in their accuracy, but device memory is not growing at the same rate. This creates fundamental challenges for training neural networks within limited memory environments. In this work, we propose ActNN, a memory-efficient training framework that stores randomly quantized activations for back propagation. We prove the convergence of ActNN for general network architectures, and we characterize the impact of quantization on the convergence via an exact expression for the gradient variance. Using our theory, we propose novel mixed-precision quantization strategies that exploit the activation's heterogeneity across feature dimensions, samples, and layers. These techniques can be readily applied to existing dynamic graph frameworks, such as PyTorch, simply by substituting the layers. We evaluate ActNN on mainstream computer vision models for classification, detection, and segmentation tasks. On all these tasks, ActNN compresses the activation to 2 bits on average, with negligible accuracy loss. ActNN reduces the memory footprint of the activation by 12x, and it enables training with a 6.6x to 14x larger batch size.

Decentralized Vehicle Coordination: The Berkeley DeepDrive Drone Dataset

2022-01-01
Fangyu Wu , Dequan Wang , Minjune Hwang +6 more
Decentralized multiagent planning has been an important field of research in robotics. An interesting and impactful application in the field is decentralized vehicle coordination in understructured road environments. For example, … Decentralized multiagent planning has been an important field of research in robotics. An interesting and impactful application in the field is decentralized vehicle coordination in understructured road environments. For example, in an intersection, it is useful yet difficult to deconflict multiple vehicles of intersecting paths in absence of a central coordinator. We learn from common sense that, for a vehicle to navigate through such understructured environments, the driver must understand and conform to the implicit "social etiquette" observed by nearby drivers. To study this implicit driving protocol, we collect the Berkeley DeepDrive Drone dataset. The dataset contains 1) a set of aerial videos recording understructured driving, 2) a collection of images and annotations to train vehicle detection models, and 3) a kit of development scripts for illustrating typical usages. We believe that the dataset is of primary interest for studying decentralized multiagent planning employed by human drivers and, of secondary interest, for computer vision in remote sensing settings.
View PDF Chat

BayesDiff: Estimating Pixel-wise Uncertainty in Diffusion via Bayesian Inference

2023-01-01
Siqi Kou , Lei Gan , Dequan Wang +2 more
Diffusion models have impressive image generation capability, but low-quality generations still exist, and their identification remains challenging due to the lack of a proper sample-wise metric. To address this, we … Diffusion models have impressive image generation capability, but low-quality generations still exist, and their identification remains challenging due to the lack of a proper sample-wise metric. To address this, we propose BayesDiff, a pixel-wise uncertainty estimator for generations from diffusion models based on Bayesian inference. In particular, we derive a novel uncertainty iteration principle to characterize the uncertainty dynamics in diffusion, and leverage the last-layer Laplace approximation for efficient Bayesian inference. The estimated pixel-wise uncertainty can not only be aggregated into a sample-wise metric to filter out low-fidelity images but also aids in augmenting successful generations and rectifying artifacts in failed generations in text-to-image tasks. Extensive experiments demonstrate the efficacy of BayesDiff and its promise for practical applications.

Iterative Object and Part Transfer for Fine-Grained Recognition

2017-01-01
Zhiqiang Shen , Yu–Gang Jiang , Dequan Wang +1 more
The aim of fine-grained recognition is to identify sub-ordinate categories in images like different species of birds. Existing works have confirmed that, in order to capture the subtle differences across … The aim of fine-grained recognition is to identify sub-ordinate categories in images like different species of birds. Existing works have confirmed that, in order to capture the subtle differences across the categories, automatic localization of objects and parts is critical. Most approaches for object and part localization relied on the bottom-up pipeline, where thousands of region proposals are generated and then filtered by pre-trained object/part models. This is computationally expensive and not scalable once the number of objects/parts becomes large. In this paper, we propose a nonparametric data-driven method for object and part localization. Given an unlabeled test image, our approach transfers annotations from a few similar images retrieved in the training set. In particular, we propose an iterative transfer strategy that gradually refine the predicted bounding boxes. Based on the located objects and parts, deep convolutional features are extracted for recognition. We evaluate our approach on the widely-used CUB200-2011 dataset and a new and large dataset called Birdsnap. On both datasets, we achieve better results than many state-of-the-art approaches, including a few using oracle (manually annotated) bounding boxes in the test images.
View PDF Chat

CoDeNet: Efficient Deployment of Input-Adaptive Object Detection on Embedded FPGAs

2020-01-01
Zhen Dong , Dequan Wang , Qijing Huang +6 more
Deploying deep learning models on embedded systems has been challenging due to limited computing resources. The majority of existing work focuses on accelerating image classification, while other fundamental vision problems, … Deploying deep learning models on embedded systems has been challenging due to limited computing resources. The majority of existing work focuses on accelerating image classification, while other fundamental vision problems, such as object detection, have not been adequately addressed. Compared with image classification, detection problems are more sensitive to the spatial variance of objects, and therefore, require specialized convolutions to aggregate spatial information. To address this need, recent work introduces dynamic deformable convolution to augment regular convolutions. However, this will lead to inefficient memory accesses of inputs with existing hardware. In this work, we harness the flexibility of FPGAs to develop a novel object detection pipeline with deformable convolutions. We show the speed-accuracy tradeoffs for a set of algorithm modifications including irregular-access versus limited-range and fixed-shape. We then Co-Design a Network CoDeNet with the modified deformable convolution and quantize it to 4-bit weights and 8-bit activations. With our high-efficiency implementation, our solution reaches 26.9 frames per second with a tiny model size of 0.76 MB while achieving 61.7 AP50 on the standard object detection dataset, Pascal VOC. With our higher accuracy implementation, our model gets to 67.1 AP50 on Pascal VOC with only 2.9 MB of parameters-20.9x smaller but 10% more accurate than Tiny-YOLO.
View PDF Chat

On-target Adaptation

2021-09-02
Dequan Wang , Shaoteng Liu , Sayna Ebrahimi +2 more
Domain adaptation seeks to mitigate the shift between training on the \emph{source} domain and testing on the \emph{target} domain. Most adaptation methods rely on the source data by joint optimization … Domain adaptation seeks to mitigate the shift between training on the \emph{source} domain and testing on the \emph{target} domain. Most adaptation methods rely on the source data by joint optimization over source data and target data. Source-free methods replace the source data with a source model by fine-tuning it on target. Either way, the majority of the parameter updates for the model representation and the classifier are derived from the source, and not the target. However, target accuracy is the goal, and so we argue for optimizing as much as possible on the target data. We show significant improvement by on-target adaptation, which learns the representation purely from target data while taking only the source predictions for supervision. In the long-tailed classification setting, we show further improvement by on-target class distribution learning, which learns the (im)balance of classes from target data.

Contrastive Test-Time Adaptation

2022-01-01
Dian Chen , Dequan Wang , Trevor Darrell +1 more
Test-time adaptation is a special setting of unsupervised domain adaptation where a trained model on the source domain has to adapt to the target domain without accessing source data. We … Test-time adaptation is a special setting of unsupervised domain adaptation where a trained model on the source domain has to adapt to the target domain without accessing source data. We propose a novel way to leverage self-supervised contrastive learning to facilitate target feature learning, along with an online pseudo labeling scheme with refinement that significantly denoises pseudo labels. The contrastive learning task is applied jointly with pseudo labeling, contrasting positive and negative pairs constructed similarly as MoCo but with source-initialized encoder, and excluding same-class negative pairs indicated by pseudo labels. Meanwhile, we produce pseudo labels online and refine them via soft voting among their nearest neighbors in the target feature space, enabled by maintaining a memory queue. Our method, AdaContrast, achieves state-of-the-art performance on major benchmarks while having several desirable properties compared to existing works, including memory efficiency, insensitivity to hyper-parameters, and better model calibration. Project page: sites.google.com/view/adacontrast.
View PDF Chat

On-target Adaptation

2021-01-01
Dequan Wang , Shaoteng Liu , Sayna Ebrahimi +2 more
Domain adaptation seeks to mitigate the shift between training on the \emph{source} domain and testing on the \emph{target} domain. Most adaptation methods rely on the source data by joint optimization … Domain adaptation seeks to mitigate the shift between training on the \emph{source} domain and testing on the \emph{target} domain. Most adaptation methods rely on the source data by joint optimization over source data and target data. Source-free methods replace the source data with a source model by fine-tuning it on target. Either way, the majority of the parameter updates for the model representation and the classifier are derived from the source, and not the target. However, target accuracy is the goal, and so we argue for optimizing as much as possible on the target data. We show significant improvement by on-target adaptation, which learns the representation purely from target data while taking only the source predictions for supervision. In the long-tailed classification setting, we show further improvement by on-target class distribution learning, which learns the (im)balance of classes from target data.
View PDF Chat

Text-guided Foundation Model Adaptation for Pathological Image Classification

2023-01-01
Yunkun Zhang , Jin Gao , Mu Zhou +4 more
The recent surge of foundation models in computer vision and natural language processing opens up perspectives in utilizing multi-modal clinical data to train large models with strong generalizability. Yet pathological … The recent surge of foundation models in computer vision and natural language processing opens up perspectives in utilizing multi-modal clinical data to train large models with strong generalizability. Yet pathological image datasets often lack biomedical text annotation and enrichment. Guiding data-efficient image diagnosis from the use of biomedical text knowledge becomes a substantial interest. In this paper, we propose to Connect Image and Text Embeddings (CITE) to enhance pathological image classification. CITE injects text insights gained from language models pre-trained with a broad range of biomedical texts, leading to adapt foundation models towards pathological image understanding. Through extensive experiments on the PatchGastric stomach tumor pathological image dataset, we demonstrate that CITE achieves leading performance compared with various baselines especially when training data is scarce. CITE offers insights into leveraging in-domain text knowledge to reinforce data-efficient pathological image classification. Code is available at https://github.com/Yunkun-Zhang/CITE.
View PDF Chat

An Extensible Framework for Open Heterogeneous Collaborative Perception

2024-01-01
Yifan Lu , Yue Hu , Yiqi Zhong +3 more
Collaborative perception aims to mitigate the limitations of single-agent perception, such as occlusions, by facilitating data exchange among multiple agents. However, most current works consider a homogeneous scenario where all … Collaborative perception aims to mitigate the limitations of single-agent perception, such as occlusions, by facilitating data exchange among multiple agents. However, most current works consider a homogeneous scenario where all agents use identity sensors and perception models. In reality, heterogeneous agent types may continually emerge and inevitably face a domain gap when collaborating with existing agents. In this paper, we introduce a new open heterogeneous problem: how to accommodate continually emerging new heterogeneous agent types into collaborative perception, while ensuring high perception performance and low integration cost? To address this problem, we propose HEterogeneous ALliance (HEAL), a novel extensible collaborative perception framework. HEAL first establishes a unified feature space with initial agents via a novel multi-scale foreground-aware Pyramid Fusion network. When heterogeneous new agents emerge with previously unseen modalities or models, we align them to the established unified space with an innovative backward alignment. This step only involves individual training on the new agent type, thus presenting extremely low training costs and high extensibility. To enrich agents' data heterogeneity, we bring OPV2V-H, a new large-scale dataset with more diverse sensor types. Extensive experiments on OPV2V-H and DAIR-V2X datasets show that HEAL surpasses SOTA methods in performance while reducing the training parameters by 91.5% when integrating 3 new agent types. We further implement a comprehensive codebase at: https://github.com/yifanlu0227/HEAL
View PDF Chat

mmPlace: Robust Place Recognition with Intermediate Frequency Signal of Low-cost Single-chip Millimeter Wave Radar

2024-03-07
Chengzhen Meng , Yifan Duan , Chenming He +3 more
Place recognition is crucial for tasks like loop-closure detection and re-localization. Single-chip millimeter wave radar (single-chip radar in short) emerges as a low-cost sensor option for place recognition, with the … Place recognition is crucial for tasks like loop-closure detection and re-localization. Single-chip millimeter wave radar (single-chip radar in short) emerges as a low-cost sensor option for place recognition, with the advantage of insensitivity to degraded visual environments. However, it encounters two challenges. Firstly, sparse point cloud from single-chip radar leads to poor performance when using current place recognition methods, which assume much denser data. Secondly, its performance significantly declines in scenarios involving rotational and lateral variations, due to limited overlap in its field of view (FOV). We propose mmPlace, a robust place recognition system to address these challenges. Specifically, mmPlace transforms intermediate frequency (IF) signal into range azimuth heatmap and employs a spatial encoder to extract features. Additionally, to improve the performance in scenarios involving rotational and lateral variations, mmPlace employs a rotating platform and concatenates heatmaps in a rotation cycle, effectively expanding the system's FOV. We evaluate mmPlace's performance on the milliSonic dataset, which is collected on the University of Science and Technology of China (USTC) campus, the city roads surrounding the campus, and an underground parking garage. The results demonstrate that mmPlace outperforms point cloud-based methods and achieves 87.37% recall@1 in scenarios involving rotational and lateral variations.
View PDF Chat

Lost in Translation: Latent Concept Misalignment in Text-to-Image Diffusion Models

2024-07-31
Juntu Zhao , Junyu Deng , Yixin Ye +3 more
Advancements in text-to-image diffusion models have broadened extensive downstream practical applications, but such models often encounter misalignment issues between text and image. Taking the generation of a combination of two … Advancements in text-to-image diffusion models have broadened extensive downstream practical applications, but such models often encounter misalignment issues between text and image. Taking the generation of a combination of two disentangled concepts as an example, say given the prompt "a tea cup of iced coke", existing models usually generate a glass cup of iced coke because the iced coke usually co-occurs with the glass cup instead of the tea one during model training. The root of such misalignment is attributed to the confusion in the latent semantic space of text-to-image diffusion models, and hence we refer to the "a tea cup of iced coke" phenomenon as Latent Concept Misalignment (LC-Mis). We leverage large language models (LLMs) to thoroughly investigate the scope of LC-Mis, and develop an automated pipeline for aligning the latent semantics of diffusion models to text prompts. Empirical assessments confirm the effectiveness of our approach, substantially reducing LC-Mis errors and enhancing the robustness and versatility of text-to-image diffusion models. Our code and dataset have been available online for reference.
View PDF Chat

Lost in Translation: Latent Concept Misalignment in Text-to-Image Diffusion Models

2024-07-31
Juntu Zhao , Junyu Deng , Yixin Ye +3 more
Advancements in text-to-image diffusion models have broadened extensive downstream practical applications, but such models often encounter misalignment issues between text and image. Taking the generation of a combination of two … Advancements in text-to-image diffusion models have broadened extensive downstream practical applications, but such models often encounter misalignment issues between text and image. Taking the generation of a combination of two disentangled concepts as an example, say given the prompt "a tea cup of iced coke", existing models usually generate a glass cup of iced coke because the iced coke usually co-occurs with the glass cup instead of the tea one during model training. The root of such misalignment is attributed to the confusion in the latent semantic space of text-to-image diffusion models, and hence we refer to the "a tea cup of iced coke" phenomenon as Latent Concept Misalignment (LC-Mis). We leverage large language models (LLMs) to thoroughly investigate the scope of LC-Mis, and develop an automated pipeline for aligning the latent semantics of diffusion models to text prompts. Empirical assessments confirm the effectiveness of our approach, substantially reducing LC-Mis errors and enhancing the robustness and versatility of text-to-image diffusion models. Our code and dataset have been available online for reference.
View PDF Chat

mmPlace: Robust Place Recognition With Intermediate Frequency Signal of Low-Cost Single-Chip Millimeter Wave Radar

2024-03-14
Chengzhen Meng , Yifan Duan , Chenming He +3 more
Place recognition is crucial for tasks like loop-closure detection and re-localization. Single-chip millimeter wave radar (single-chip radar in short) emerges as a low-cost sensor option for place recognition, with the … Place recognition is crucial for tasks like loop-closure detection and re-localization. Single-chip millimeter wave radar (single-chip radar in short) emerges as a low-cost sensor option for place recognition, with the advantage of insensitivity to degraded visual environments. However, it encounters two challenges. Firstly, sparse point cloud from single-chip radar leads to poor performance when using current place recognition methods, which assume much denser data. Secondly, its performance significantly declines in scenarios involving rotational and lateral variations, due to limited overlap in its field of view (FOV). We propose mmPlace, a robust place recognition system to address these challenges. Specifically, mmPlace transforms intermediate frequency (IF) signal into range azimuth heatmap and employs a spatial encoder to extract features. Additionally, to improve the performance in scenarios involving rotational and lateral variations, mmPlace employs a rotating platform and concatenates heatmaps in a rotation cycle, effectively expanding the system's FOV. We evaluate mmPlace's performance on the milliSonic dataset, which is collected on the University of Science and Technology of China (USTC) campus, the city roads surrounding the campus, and an underground parking garage. The results demonstrate that mmPlace outperforms point cloud-based methods and achieves 87.37% recall@1 in scenarios involving rotational and lateral variations.
View PDF Chat

mmPlace: Robust Place Recognition with Intermediate Frequency Signal of Low-cost Single-chip Millimeter Wave Radar

2024-03-07
Chengzhen Meng , Yifan Duan , Chenming He +3 more
Place recognition is crucial for tasks like loop-closure detection and re-localization. Single-chip millimeter wave radar (single-chip radar in short) emerges as a low-cost sensor option for place recognition, with the … Place recognition is crucial for tasks like loop-closure detection and re-localization. Single-chip millimeter wave radar (single-chip radar in short) emerges as a low-cost sensor option for place recognition, with the advantage of insensitivity to degraded visual environments. However, it encounters two challenges. Firstly, sparse point cloud from single-chip radar leads to poor performance when using current place recognition methods, which assume much denser data. Secondly, its performance significantly declines in scenarios involving rotational and lateral variations, due to limited overlap in its field of view (FOV). We propose mmPlace, a robust place recognition system to address these challenges. Specifically, mmPlace transforms intermediate frequency (IF) signal into range azimuth heatmap and employs a spatial encoder to extract features. Additionally, to improve the performance in scenarios involving rotational and lateral variations, mmPlace employs a rotating platform and concatenates heatmaps in a rotation cycle, effectively expanding the system's FOV. We evaluate mmPlace's performance on the milliSonic dataset, which is collected on the University of Science and Technology of China (USTC) campus, the city roads surrounding the campus, and an underground parking garage. The results demonstrate that mmPlace outperforms point cloud-based methods and achieves 87.37% recall@1 in scenarios involving rotational and lateral variations.
View PDF Chat

Data-Centric Foundation Models in Computational Healthcare: A Survey

2024-01-01
Yunkun Zhang , Jin Gao , Zheling Tan +5 more
The advent of foundation models (FMs) as an emerging suite of AI techniques has struck a wave of opportunities in computational healthcare. The interactive nature of these models, guided by … The advent of foundation models (FMs) as an emerging suite of AI techniques has struck a wave of opportunities in computational healthcare. The interactive nature of these models, guided by pre-training data and human instructions, has ignited a data-centric AI paradigm that emphasizes better data characterization, quality, and scale. In healthcare AI, obtaining and processing high-quality clinical data records has been a longstanding challenge, ranging from data quantity, annotation, patient privacy, and ethics. In this survey, we investigate a wide range of data-centric approaches in the FM era (from model pre-training to inference) towards improving the healthcare workflow. We discuss key perspectives in AI security, assessment, and alignment with human values. Finally, we offer a promising outlook of FM-based analytics to enhance the performance of patient outcome and clinical workflow in the evolving landscape of healthcare and medicine. We provide an up-to-date list of healthcare-related foundation models and datasets at https://github.com/Yunkun-Zhang/Data-Centric-FM-Healthcare .

An Extensible Framework for Open Heterogeneous Collaborative Perception

2024-01-01
Yifan Lu , Yue Hu , Yiqi Zhong +3 more
Collaborative perception aims to mitigate the limitations of single-agent perception, such as occlusions, by facilitating data exchange among multiple agents. However, most current works consider a homogeneous scenario where all … Collaborative perception aims to mitigate the limitations of single-agent perception, such as occlusions, by facilitating data exchange among multiple agents. However, most current works consider a homogeneous scenario where all agents use identity sensors and perception models. In reality, heterogeneous agent types may continually emerge and inevitably face a domain gap when collaborating with existing agents. In this paper, we introduce a new open heterogeneous problem: how to accommodate continually emerging new heterogeneous agent types into collaborative perception, while ensuring high perception performance and low integration cost? To address this problem, we propose HEterogeneous ALliance (HEAL), a novel extensible collaborative perception framework. HEAL first establishes a unified feature space with initial agents via a novel multi-scale foreground-aware Pyramid Fusion network. When heterogeneous new agents emerge with previously unseen modalities or models, we align them to the established unified space with an innovative backward alignment. This step only involves individual training on the new agent type, thus presenting extremely low training costs and high extensibility. To enrich agents' data heterogeneity, we bring OPV2V-H, a new large-scale dataset with more diverse sensor types. Extensive experiments on OPV2V-H and DAIR-V2X datasets show that HEAL surpasses SOTA methods in performance while reducing the training parameters by 91.5% when integrating 3 new agent types. We further implement a comprehensive codebase at: https://github.com/yifanlu0227/HEAL
View PDF Chat

Towards General Purpose Medical AI: Continual Learning Medical Foundation Model

2023-01-01
Huahui Yi , Ziyuan Qin , Qicheng Lao +5 more
Inevitable domain and task discrepancies in real-world scenarios can impair the generalization performance of the pre-trained deep models for medical data. Therefore, we audaciously propose that we should build a … Inevitable domain and task discrepancies in real-world scenarios can impair the generalization performance of the pre-trained deep models for medical data. Therefore, we audaciously propose that we should build a general-purpose medical AI system that can be seamlessly adapted to downstream domains/tasks. Since the domain/task adaption procedures usually involve additional labeling work for the target data, designing a data-efficient adaption algorithm is desired to save the cost of transferring the learned knowledge. Our recent work found that vision-language models (VLMs) are efficient learners with extraordinary cross-domain ability. Therefore, in this work, we further explore the possibility of leveraging pre-trained VLMs as medical foundation models for building general-purpose medical AI, where we thoroughly investigate three machine-learning paradigms, i.e., domain/task-specialized learning, joint learning, and continual learning, for training the VLMs and evaluate their generalization performance on cross-domain and cross-task test sets. To alleviate the catastrophic forgetting during sequential training, we employ rehearsal learning and receive a sharp boost in terms of generalization capability. In a nutshell, our empirical evidence suggests that continual learning may be a practical and efficient learning paradigm for the medical foundation model. And we hope researchers can use our empirical evidence as basement to further explore the path toward medical foundation model.

Text-guided Foundation Model Adaptation for Pathological Image Classification

2023-01-01
Yunkun Zhang , Jin Gao , Mu Zhou +4 more
The recent surge of foundation models in computer vision and natural language processing opens up perspectives in utilizing multi-modal clinical data to train large models with strong generalizability. Yet pathological … The recent surge of foundation models in computer vision and natural language processing opens up perspectives in utilizing multi-modal clinical data to train large models with strong generalizability. Yet pathological image datasets often lack biomedical text annotation and enrichment. Guiding data-efficient image diagnosis from the use of biomedical text knowledge becomes a substantial interest. In this paper, we propose to Connect Image and Text Embeddings (CITE) to enhance pathological image classification. CITE injects text insights gained from language models pre-trained with a broad range of biomedical texts, leading to adapt foundation models towards pathological image understanding. Through extensive experiments on the PatchGastric stomach tumor pathological image dataset, we demonstrate that CITE achieves leading performance compared with various baselines especially when training data is scarce. CITE offers insights into leveraging in-domain text knowledge to reinforce data-efficient pathological image classification. Code is available at https://github.com/Yunkun-Zhang/CITE.
View PDF Chat

BayesDiff: Estimating Pixel-wise Uncertainty in Diffusion via Bayesian Inference

2023-01-01
Siqi Kou , Lei Gan , Dequan Wang +2 more
Diffusion models have impressive image generation capability, but low-quality generations still exist, and their identification remains challenging due to the lack of a proper sample-wise metric. To address this, we … Diffusion models have impressive image generation capability, but low-quality generations still exist, and their identification remains challenging due to the lack of a proper sample-wise metric. To address this, we propose BayesDiff, a pixel-wise uncertainty estimator for generations from diffusion models based on Bayesian inference. In particular, we derive a novel uncertainty iteration principle to characterize the uncertainty dynamics in diffusion, and leverage the last-layer Laplace approximation for efficient Bayesian inference. The estimated pixel-wise uncertainty can not only be aggregated into a sample-wise metric to filter out low-fidelity images but also aids in augmenting successful generations and rectifying artifacts in failed generations in text-to-image tasks. Extensive experiments demonstrate the efficacy of BayesDiff and its promise for practical applications.

Contrastive Test-Time Adaptation

2022-06-01
Dian Chen , Dequan Wang , Trevor Darrell +1 more
Test-time adaptation is a special setting of unsupervised domain adaptation where a trained model on the source domain has to adapt to the target domain without accessing source data. We … Test-time adaptation is a special setting of unsupervised domain adaptation where a trained model on the source domain has to adapt to the target domain without accessing source data. We propose a novel way to leverage self-supervised contrastive learning to facilitate target feature learning, along with an online pseudo labeling scheme with refinement that significantly denoises pseudo labels. The contrastive learning task is applied jointly with pseudo labeling, contrasting positive and negative pairs constructed similarly as MoCo but with source-initialized encoder, and excluding same-class negative pairs indicated by pseudo labels. Meanwhile, we produce pseudo labels online and refine them via soft voting among their nearest neighbors in the target feature space, enabled by maintaining a memory queue. Our method, AdaContrast, achieves state-of-the-art performance on major benchmarks while having several desirable properties compared to existing works, including memory efficiency, insensitivity to hyper-parameters, and better model calibration. Code is released at https://github.com/DianCh/AdaContrast.
View PDF Chat

Contrastive Test-Time Adaptation

2022-01-01
Dian Chen , Dequan Wang , Trevor Darrell +1 more
Test-time adaptation is a special setting of unsupervised domain adaptation where a trained model on the source domain has to adapt to the target domain without accessing source data. We … Test-time adaptation is a special setting of unsupervised domain adaptation where a trained model on the source domain has to adapt to the target domain without accessing source data. We propose a novel way to leverage self-supervised contrastive learning to facilitate target feature learning, along with an online pseudo labeling scheme with refinement that significantly denoises pseudo labels. The contrastive learning task is applied jointly with pseudo labeling, contrasting positive and negative pairs constructed similarly as MoCo but with source-initialized encoder, and excluding same-class negative pairs indicated by pseudo labels. Meanwhile, we produce pseudo labels online and refine them via soft voting among their nearest neighbors in the target feature space, enabled by maintaining a memory queue. Our method, AdaContrast, achieves state-of-the-art performance on major benchmarks while having several desirable properties compared to existing works, including memory efficiency, insensitivity to hyper-parameters, and better model calibration. Project page: sites.google.com/view/adacontrast.
View PDF Chat

Back to the Source: Diffusion-Driven Test-Time Adaptation

2022-01-01
Jin Gao , Jialing Zhang , Xihui Liu +3 more
Test-time adaptation harnesses test inputs to improve the accuracy of a model trained on source data when tested on shifted target data. Existing methods update the source model by (re-)training … Test-time adaptation harnesses test inputs to improve the accuracy of a model trained on source data when tested on shifted target data. Existing methods update the source model by (re-)training on each target domain. While effective, re-training is sensitive to the amount and order of the data and the hyperparameters for optimization. We instead update the target data, by projecting all test inputs toward the source domain with a generative diffusion model. Our diffusion-driven adaptation method, DDA, shares its models for classification and generation across all domains. Both models are trained on the source domain, then fixed during testing. We augment diffusion with image guidance and self-ensembling to automatically decide how much to adapt. Input adaptation by DDA is more robust than prior model adaptation approaches across a variety of corruptions, architectures, and data regimes on the ImageNet-C benchmark. With its input-wise updates, DDA succeeds where model adaptation degrades on too little data in small batches, dependent data in non-uniform order, or mixed data with multiple corruptions.
View PDF Chat

GACT: Activation Compressed Training for Generic Network Architectures

2022-01-01
Xiaoxuan Liu , Lianmin Zheng , Dequan Wang +7 more
Training large neural network (NN) models requires extensive memory resources, and Activation Compressed Training (ACT) is a promising approach to reduce training memory footprint. This paper presents GACT, an ACT … Training large neural network (NN) models requires extensive memory resources, and Activation Compressed Training (ACT) is a promising approach to reduce training memory footprint. This paper presents GACT, an ACT framework to support a broad range of machine learning tasks for generic NN architectures with limited domain knowledge. By analyzing a linearized version of ACT's approximate gradient, we prove the convergence of GACT without prior knowledge on operator type or model architecture. To make training stable, we propose an algorithm that decides the compression ratio for each tensor by estimating its impact on the gradient at run time. We implement GACT as a PyTorch library that readily applies to any NN architecture. GACT reduces the activation memory for convolutional NNs, transformers, and graph NNs by up to 8.1x, enabling training with a 4.2x to 24.7x larger batch size, with negligible accuracy loss. We implement GACT as a PyTorch library at https://github.com/LiuXiaoxuanPKU/GACT-ICML.
View PDF Chat

Decentralized Vehicle Coordination: The Berkeley DeepDrive Drone Dataset

2022-01-01
Fangyu Wu , Dequan Wang , Minjune Hwang +6 more
Decentralized multiagent planning has been an important field of research in robotics. An interesting and impactful application in the field is decentralized vehicle coordination in understructured road environments. For example, … Decentralized multiagent planning has been an important field of research in robotics. An interesting and impactful application in the field is decentralized vehicle coordination in understructured road environments. For example, in an intersection, it is useful yet difficult to deconflict multiple vehicles of intersecting paths in absence of a central coordinator. We learn from common sense that, for a vehicle to navigate through such understructured environments, the driver must understand and conform to the implicit "social etiquette" observed by nearby drivers. To study this implicit driving protocol, we collect the Berkeley DeepDrive Drone dataset. The dataset contains 1) a set of aerial videos recording understructured driving, 2) a collection of images and annotations to train vehicle detection models, and 3) a kit of development scripts for illustrating typical usages. We believe that the dataset is of primary interest for studying decentralized multiagent planning employed by human drivers and, of secondary interest, for computer vision in remote sensing settings.
View PDF Chat

MAROAM: Map-based Radar SLAM through Two-step Feature Selection

2022-01-01
Dequan Wang , Yifan Duan , Xiaoran Fan +3 more
In this letter, we propose MAROAM, a millimeter wave radar-based SLAM framework, which employs a two-step feature selection process to build the global consistent map. Specifically, we first extract feature … In this letter, we propose MAROAM, a millimeter wave radar-based SLAM framework, which employs a two-step feature selection process to build the global consistent map. Specifically, we first extract feature points from raw data based on their local geometric properties to filter out those points that violate the principle of millimeter-wave radar imaging. Then, we further employ another round of probabilistic feature selection by examining how often and how recent the feature point has been detected in the proceeding frames. With such a two-step feature selection, we establish a global consistent map for accurate and robust pose estimation as well as other downstream tasks. At last, we perform loop closure and graph optimization in the back-end, further reducing the accumulated drift error. We evaluate the performance of MAROAM on the three datasets: the Oxford Radar RobotCar Dataset, the MulRan Dataset and the Boreas Dataset. We consider a variety of experimental settings with different scenery, weather, and road conditions. The experimental results show that the accuracy of MAROAM is 7.95%, 37.0% and 8.9% higher than the currently best-performing algorithms on these three datasets, respectively. The ablation results also show that our map-based odometry performs 28.6% better than the commonly used scan-to-frames method. Finally, as devoted contributors to the open-source community, we will open source the algorithm after the paper is accepted.
View PDF Chat

On-target Adaptation

2021-09-02
Dequan Wang , Shaoteng Liu , Sayna Ebrahimi +2 more
Domain adaptation seeks to mitigate the shift between training on the \emph{source} domain and testing on the \emph{target} domain. Most adaptation methods rely on the source data by joint optimization … Domain adaptation seeks to mitigate the shift between training on the \emph{source} domain and testing on the \emph{target} domain. Most adaptation methods rely on the source data by joint optimization over source data and target data. Source-free methods replace the source data with a source model by fine-tuning it on target. Either way, the majority of the parameter updates for the model representation and the classifier are derived from the source, and not the target. However, target accuracy is the goal, and so we argue for optimizing as much as possible on the target data. We show significant improvement by on-target adaptation, which learns the representation purely from target data while taking only the source predictions for supervision. In the long-tailed classification setting, we show further improvement by on-target class distribution learning, which learns the (im)balance of classes from target data.

CoDeNet: Efficient Deployment of Input-Adaptive Object Detection on Embedded FPGAs

2021-02-17
Qijing Huang , Dequan Wang , Zhen Dong +6 more
Deploying deep learning models on embedded systems for computer vision tasks has been challenging due to limited compute resources and strict energy budgets. The majority of existing work focuses on … Deploying deep learning models on embedded systems for computer vision tasks has been challenging due to limited compute resources and strict energy budgets. The majority of existing work focuses on accelerating image classification, while other fundamental vision problems, such as object detection, have not been adequately addressed. Compared with image classification, detection problems are more sensitive to the spatial variance of objects, and therefore, require specialized convolutions to aggregate spatial information. To address this need, recent work introduces dynamic deformable convolution to augment regular convolutions. Regular convolutions process a fixed grid of pixels across all the spatial locations in an image, while dynamic deformable convolution may access arbitrary pixels in the image with the access pattern being input-dependent and varying with spatial location. These properties lead to inefficient memory accesses of inputs with existing hardware.
View PDF Chat

ActNN: Reducing Training Memory Footprint via 2-Bit Activation Compressed Training

2021-01-01
Jianfei Chen , Lianmin Zheng , Zhewei Yao +4 more
The increasing size of neural network models has been critical for improvements in their accuracy, but device memory is not growing at the same rate. This creates fundamental challenges for … The increasing size of neural network models has been critical for improvements in their accuracy, but device memory is not growing at the same rate. This creates fundamental challenges for training neural networks within limited memory environments. In this work, we propose ActNN, a memory-efficient training framework that stores randomly quantized activations for back propagation. We prove the convergence of ActNN for general network architectures, and we characterize the impact of quantization on the convergence via an exact expression for the gradient variance. Using our theory, we propose novel mixed-precision quantization strategies that exploit the activation's heterogeneity across feature dimensions, samples, and layers. These techniques can be readily applied to existing dynamic graph frameworks, such as PyTorch, simply by substituting the layers. We evaluate ActNN on mainstream computer vision models for classification, detection, and segmentation tasks. On all these tasks, ActNN compresses the activation to 2 bits on average, with negligible accuracy loss. ActNN reduces the memory footprint of the activation by 12x, and it enables training with a 6.6x to 14x larger batch size.

Fighting Gradients with Gradients: Dynamic Defenses against Adversarial Attacks

2021-01-01
Dequan Wang , An Ju , Evan Shelhamer +2 more
Adversarial attacks optimize against models to defeat defenses. Existing defenses are static, and stay the same once trained, even while attacks change. We argue that models should fight back, and … Adversarial attacks optimize against models to defeat defenses. Existing defenses are static, and stay the same once trained, even while attacks change. We argue that models should fight back, and optimize their defenses against attacks at test time. We propose dynamic defenses, to adapt the model and input during testing, by defensive entropy minimization (dent). Dent alters testing, but not training, for compatibility with existing models and train-time defenses. Dent improves the robustness of adversarially-trained defenses and nominally-trained models against white-box, black-box, and adaptive attacks on CIFAR-10/100 and ImageNet. In particular, dent boosts state-of-the-art defenses by 20+ points absolute against AutoAttack on CIFAR-10 at $\epsilon_\infty$ = 8/255.
View PDF Chat

On-target Adaptation

2021-01-01
Dequan Wang , Shaoteng Liu , Sayna Ebrahimi +2 more
Domain adaptation seeks to mitigate the shift between training on the \emph{source} domain and testing on the \emph{target} domain. Most adaptation methods rely on the source data by joint optimization … Domain adaptation seeks to mitigate the shift between training on the \emph{source} domain and testing on the \emph{target} domain. Most adaptation methods rely on the source data by joint optimization over source data and target data. Source-free methods replace the source data with a source model by fine-tuning it on target. Either way, the majority of the parameter updates for the model representation and the classifier are derived from the source, and not the target. However, target accuracy is the goal, and so we argue for optimizing as much as possible on the target data. We show significant improvement by on-target adaptation, which learns the representation purely from target data while taking only the source predictions for supervision. In the long-tailed classification setting, we show further improvement by on-target class distribution learning, which learns the (im)balance of classes from target data.
View PDF Chat

Fully Test-time Adaptation by Entropy Minimization

2020-06-18
Dequan Wang , Evan Shelhamer , Shaoteng Liu +2 more
A model must adapt itself to generalize to new and different data during testing. In this setting of fully test-time adaptation the model has only the test data and its … A model must adapt itself to generalize to new and different data during testing. In this setting of fully test-time adaptation the model has only the test data and its own parameters. We propose to adapt by test entropy minimization (tent): we optimize the model for confidence as measured by the entropy of its predictions. Our method estimates normalization statistics and optimizes channel-wise affine transformations to update online on each batch. Tent reduces generalization error for image classification on corrupted ImageNet and CIFAR-10/100 and reaches a new state-of-the-art error on ImageNet-C. Tent handles source-free domain adaptation on digit recognition from SVHN to MNIST/MNIST-M/USPS, on semantic segmentation from GTA to Cityscapes, and on the VisDA-C benchmark. These results are achieved in one epoch of test-time optimization without altering training.

Tent: Fully Test-time Adaptation by Entropy Minimization

2020-06-18
Dequan Wang , Evan Shelhamer , Shaoteng Liu +2 more
A model must adapt itself to generalize to new and different data during testing. In this setting of fully test-time adaptation the model has only the test data and its … A model must adapt itself to generalize to new and different data during testing. In this setting of fully test-time adaptation the model has only the test data and its own parameters. We propose to adapt by test entropy minimization (tent): we optimize the model for confidence as measured by the entropy of its predictions. Our method estimates normalization statistics and optimizes channel-wise affine transformations to update online on each batch. Tent reduces generalization error for image classification on corrupted ImageNet and CIFAR-10/100 and reaches a new state-of-the-art error on ImageNet-C. Tent handles source-free domain adaptation on digit recognition from SVHN to MNIST/MNIST-M/USPS, on semantic segmentation from GTA to Cityscapes, and on the VisDA-C benchmark. These results are achieved in one epoch of test-time optimization without altering training.
View PDF Chat

CoDeNet: Algorithm-hardware Co-design for Deformable Convolution

2020-06-12
Zhen Dong , Dequan Wang , Qijing Huang +5 more
Deploying deep learning models on embedded systems for computer vision tasks has been challenging due to limited compute resources and strict energy budgets. The majority of existing work focuses on … Deploying deep learning models on embedded systems for computer vision tasks has been challenging due to limited compute resources and strict energy budgets. The majority of existing work focuses on accelerating image classification, while other fundamental vision problems, such as object detection, have not been adequately addressed. Compared with image classification, detection problems are more sensitive to the spatial variance of objects, and therefore, require specialized convolutions to aggregate spatial information. To address this, recent work proposes dynamic deformable convolution to augment regular convolutions. Regular convolutions process a fixed grid of pixels across all the spatial locations in an image, while dynamic deformable convolution may access arbitrary pixels in the image and the access pattern is input-dependent and varies per spatial location. These properties lead to inefficient memory accesses of inputs with existing hardware. In this work, we first investigate the overhead of the deformable convolution on embedded FPGA SoCs, and introduce a depthwise deformable convolution to reduce the total number of operations required. We then show the speed-accuracy tradeoffs for a set of algorithm modifications including irregular-access versus limited-range and fixed-shape. We evaluate these algorithmic changes with corresponding hardware optimizations. Results show a 1.36x and 9.76x speedup respectively for the full and depthwise deformable convolution on the embedded FPGA accelerator with minor accuracy loss on the object detection task. We then co-design an efficient network CoDeNet with the modified deformable convolution for object detection and quantize the network to 4-bit weights and 8-bit activations. Results show that our designs lie on the pareto-optimal front of the latency-accuracy tradeoff for the object detection task on embedded FPGAs

Algorithm-hardware Co-design for Deformable Convolution

2020-01-01
Qijing Huang , Dequan Wang , Yizhao Gao +5 more
FPGAs provide a flexible and efficient platform to accelerate rapidly-changing algorithms for computer vision. The majority of existing work focuses on accelerating image classification, while other fundamental vision problems, including … FPGAs provide a flexible and efficient platform to accelerate rapidly-changing algorithms for computer vision. The majority of existing work focuses on accelerating image classification, while other fundamental vision problems, including object detection and instance segmentation, have not been adequately addressed. Compared with image classification, detection problems are more sensitive to the spatial variance of objects, and therefore, require specialized convolutions to aggregate spatial information. To address this, recent work proposes dynamic deformable convolution to augment regular convolutions. Regular convolutions process a fixed grid of pixels across all the spatial locations in an image, while dynamic deformable convolutions may access arbitrary pixels in the image and the access pattern is input-dependent and varies per spatial location. These properties lead to inefficient memory accesses of inputs with existing hardware. In this work, we first investigate the overhead of the deformable convolution on embedded FPGA SoCs, and then show the accuracy-latency tradeoffs for a set of algorithm modifications including full versus depthwise, fixed-shape, and limited-range. These modifications benefit the energy efficiency for embedded devices in general as they reduce the compute complexity. We then build an efficient object detection network with modified deformable convolutions and quantize the network using state-of-the-art quantization methods. We implement a unified hardware engine on FPGA to support all the operations in the network. Preliminary experiments show that little accuracy is compromised and speedup can be achieved with our co-design optimization for the deformable convolution.
View PDF Chat

BEV-Seg: Bird's Eye View Semantic Segmentation Using Geometry and Semantic Point Cloud

2020-01-01
Mong H. Ng , Kaahan Radia , Chen Jian-fei +3 more
Bird's-eye-view (BEV) is a powerful and widely adopted representation for road scenes that captures surrounding objects and their spatial locations, along with overall context in the scene. In this work, … Bird's-eye-view (BEV) is a powerful and widely adopted representation for road scenes that captures surrounding objects and their spatial locations, along with overall context in the scene. In this work, we focus on bird's eye semantic segmentation, a task that predicts pixel-wise semantic segmentation in BEV from side RGB images. This task is made possible by simulators such as Carla, which allow for cheap data collection, arbitrary camera placements, and supervision in ways otherwise not possible in the real world. There are two main challenges to this task: the view transformation from side view to bird's eye view, as well as transfer learning to unseen domains. Existing work transforms between views through fully connected layers and transfer learns via GANs. This suffers from a lack of depth reasoning and performance degradation across domains. Our novel 2-staged perception pipeline explicitly predicts pixel depths and combines them with pixel semantics in an efficient manner, allowing the model to leverage depth information to infer objects' spatial locations in the BEV. In addition, we transfer learning by abstracting high-level geometric features and predicting an intermediate representation that is common across different domains. We publish a new dataset called BEVSEG-Carla and show that our approach improves state-of-the-art by 24% mIoU and performs well when transferred to a new domain.

Tent: Fully Test-time Adaptation by Entropy Minimization

2020-01-01
Dequan Wang , Evan Shelhamer , Shaoteng Liu +2 more
A model must adapt itself to generalize to new and different data during testing. In this setting of fully test-time adaptation the model has only the test data and its … A model must adapt itself to generalize to new and different data during testing. In this setting of fully test-time adaptation the model has only the test data and its own parameters. We propose to adapt by test entropy minimization (tent): we optimize the model for confidence as measured by the entropy of its predictions. Our method estimates normalization statistics and optimizes channel-wise affine transformations to update online on each batch. Tent reduces generalization error for image classification on corrupted ImageNet and CIFAR-10/100 and reaches a new state-of-the-art error on ImageNet-C. Tent handles source-free domain adaptation on digit recognition from SVHN to MNIST/MNIST-M/USPS, on semantic segmentation from GTA to Cityscapes, and on the VisDA-C benchmark. These results are achieved in one epoch of test-time optimization without altering training.

CoDeNet: Efficient Deployment of Input-Adaptive Object Detection on Embedded FPGAs

2020-01-01
Zhen Dong , Dequan Wang , Qijing Huang +6 more
Deploying deep learning models on embedded systems has been challenging due to limited computing resources. The majority of existing work focuses on accelerating image classification, while other fundamental vision problems, … Deploying deep learning models on embedded systems has been challenging due to limited computing resources. The majority of existing work focuses on accelerating image classification, while other fundamental vision problems, such as object detection, have not been adequately addressed. Compared with image classification, detection problems are more sensitive to the spatial variance of objects, and therefore, require specialized convolutions to aggregate spatial information. To address this need, recent work introduces dynamic deformable convolution to augment regular convolutions. However, this will lead to inefficient memory accesses of inputs with existing hardware. In this work, we harness the flexibility of FPGAs to develop a novel object detection pipeline with deformable convolutions. We show the speed-accuracy tradeoffs for a set of algorithm modifications including irregular-access versus limited-range and fixed-shape. We then Co-Design a Network CoDeNet with the modified deformable convolution and quantize it to 4-bit weights and 8-bit activations. With our high-efficiency implementation, our solution reaches 26.9 frames per second with a tiny model size of 0.76 MB while achieving 61.7 AP50 on the standard object detection dataset, Pascal VOC. With our higher accuracy implementation, our model gets to 67.1 AP50 on Pascal VOC with only 2.9 MB of parameters-20.9x smaller but 10% more accurate than Tiny-YOLO.
View PDF Chat

Algorithm-hardware Co-design for Deformable Convolution

2019-12-01
Qijing Huang , Dequan Wang , Yizhao Gao +5 more
FPGAs provide a flexible and efficient platform to accelerate rapidly-changing algorithms for computer vision. The majority of existing work focuses on accelerating image classification, while other fundamental vision problems, including … FPGAs provide a flexible and efficient platform to accelerate rapidly-changing algorithms for computer vision. The majority of existing work focuses on accelerating image classification, while other fundamental vision problems, including object detection and instance segmentation, have not been adequately addressed. Compared with image classification, detection problems are more sensitive to the spatial variance of objects, and therefore, require specialized convolutions to aggregate spatial information. To address this, recent work proposes dynamic deformable convolution to augment regular convolutions. Regular convolutions process a fixed grid of pixels across all the spatial locations in an image, while dynamic deformable convolutions may access arbitrary pixels. The access pattern of deformable convolutions is input-dependent and varies per spatial location. These properties lead to inefficient memory accesses of inputs with existing hardware. In this work, we first investigate the overhead of the deformable convolution on embedded FPGA SoCs, and then show the accuracy-latency tradeoffs for a set of algorithm modifications, including full versus depthwise, fixed-shape, and limited-range. These modifications benefit the efficiency of the embedded accelerator in general. We build an efficient object detection network with modified deformable convolutions and quantize the network using state-of-the-art quantization methods. Experiments show that our co-design optimization for the deformable convolution achieves significant hardware speedup with little accuracy compromised.
View PDF Chat

Monocular Plan View Networks for Autonomous Driving

2019-11-01
Dequan Wang , Coline Devin , Qi-Zhi Cai +2 more
Convolutions on monocular dash cam videos capture spatial invariances in the image plane but do not explicitly reason about distances and depth. We propose a simple transformation of observations into … Convolutions on monocular dash cam videos capture spatial invariances in the image plane but do not explicitly reason about distances and depth. We propose a simple transformation of observations into a bird's eye view, also known as plan view, for end-to-end control. We detect vehicles and pedestrians in the first person view and project them into an overhead plan view. This representation provides an abstraction of the environment from which a deep network can easily deduce the positions and directions of entities. Additionally, the plan view enables us to leverage advances in 3D object detection in conjunction with deep policy learning. We evaluate our monocular plan view network on the photo-realistic Grand Theft Auto V simulator. A network using both a plan view and front view causes less than half as many collisions as previous detection-based methods and an order of magnitude fewer collisions than pure pixel-based policies.
View PDF Chat

Joint Monocular 3D Vehicle Detection and Tracking

2019-10-01
Hou-Ning Hu , Qi-Zhi Cai , Dequan Wang +5 more
Vehicle 3D extents and trajectories are critical cues for predicting the future location of vehicles and planning future agent ego-motion based on those predictions. In this paper, we propose a … Vehicle 3D extents and trajectories are critical cues for predicting the future location of vehicles and planning future agent ego-motion based on those predictions. In this paper, we propose a novel online framework for 3D vehicle detection and tracking from monocular videos. The framework can not only associate detections of vehicles in motion over time, but also estimate their complete 3D bounding box information from a sequence of 2D images captured on a moving platform. Our method leverages 3D box depth-ordering matching for robust instance association and utilizes 3D trajectory prediction for re-identification of occluded vehicles. We also design a motion learning module based on an LSTM for more accurate long-term motion extrapolation. Our experiments on simulation, KITTI, and Argoverse datasets show that our 3D tracking pipeline offers robust data association and tracking. On Argoverse, our image-based method is significantly better for tracking 3D vehicles within 30 meters than the LiDAR-centric baseline methods.
View PDF Chat

Monocular Plan View Networks for Autonomous Driving

2019-05-16
Dequan Wang , Coline Devin , Qi-Zhi Cai +2 more
Convolutions on monocular dash cam videos capture spatial invariances in the image plane but do not explicitly reason about distances and depth. We propose a simple transformation of observations into … Convolutions on monocular dash cam videos capture spatial invariances in the image plane but do not explicitly reason about distances and depth. We propose a simple transformation of observations into a bird's eye view, also known as plan view, for end-to-end control. We detect vehicles and pedestrians in the first person view and project them into an overhead plan view. This representation provides an abstraction of the environment from which a deep network can easily deduce the positions and directions of entities. Additionally, the plan view enables us to leverage advances in 3D object detection in conjunction with deep policy learning. We evaluate our monocular plan view network on the photo-realistic Grand Theft Auto V simulator. A network using both a plan view and front view causes less than half as many collisions as previous detection-based methods and an order of magnitude fewer collisions than pure pixel-based policies.

Deep Object-Centric Policies for Autonomous Driving

2019-05-01
Dequan Wang , Coline Devin , Qi-Zhi Cai +2 more
While learning visuomotor skills in an end-to-end manner is appealing, deep neural networks are often uninterpretable and fail in surprising ways. For robotics tasks, such as autonomous driving, models that … While learning visuomotor skills in an end-to-end manner is appealing, deep neural networks are often uninterpretable and fail in surprising ways. For robotics tasks, such as autonomous driving, models that explicitly represent objects may be more robust to new scenes and provide intuitive visualizations. We describe a taxonomy of “object-centric” models which leverage both object instances and end-to-end learning. In the Grand Theft Auto V simulator, we show that object-centric models outperform object-agnostic methods in scenes with other vehicles and pedestrians, even with an imperfect detector. We also demonstrate that our architectures perform well on real-world environments by evaluating on the Berkeley DeepDrive Video dataset, where an object-centric model outperforms object-agnostic models in the low-data regimes.
View PDF Chat

Blurring the Line Between Structure and Learning to Optimize and Adapt Receptive Fields.

2019-04-25
Evan Shelhamer , Dequan Wang , Trevor Darrell
The visual world is vast and varied, but its variations divide into structured and unstructured factors. We compose free-form filters and structured Gaussian filters, optimized end-to-end, to factorize deep representations … The visual world is vast and varied, but its variations divide into structured and unstructured factors. We compose free-form filters and structured Gaussian filters, optimized end-to-end, to factorize deep representations and learn both local features and their degree of locality. Our semi-structured composition is strictly more expressive than free-form filtering, and changes in its structured parameters would require changes in free-form architecture. In effect this optimizes over receptive field size and shape, tuning locality to the data and task. Dynamic inference, in which the Gaussian structure varies with the input, adapts receptive field size to compensate for local scale variation. Optimizing receptive field size improves semantic segmentation accuracy on Cityscapes by 1-2 points for strong dilated and skip architectures and by up to 10 points for suboptimal designs. Adapting receptive fields by dynamic Gaussian structure further improves results, equaling the accuracy of free-form deformation while improving efficiency.

Objects as Points

2019-04-16
Xingyi Zhou , Dequan Wang , Philipp Krähenbühl
Detection identifies objects as axis-aligned boxes in an image. Most successful object detectors enumerate a nearly exhaustive list of potential object locations and classify each. This is wasteful, inefficient, and … Detection identifies objects as axis-aligned boxes in an image. Most successful object detectors enumerate a nearly exhaustive list of potential object locations and classify each. This is wasteful, inefficient, and requires additional post-processing. In this paper, we take a different approach. We model an object as a single point --- the center point of its bounding box. Our detector uses keypoint estimation to find center points and regresses to all other object properties, such as size, 3D location, orientation, and even pose. Our center point based approach, CenterNet, is end-to-end differentiable, simpler, faster, and more accurate than corresponding bounding box based detectors. CenterNet achieves the best speed-accuracy trade-off on the MS COCO dataset, with 28.1% AP at 142 FPS, 37.4% AP at 52 FPS, and 45.1% AP with multi-scale testing at 1.4 FPS. We use the same approach to estimate 3D bounding box in the KITTI benchmark and human pose on the COCO keypoint dataset. Our method performs competitively with sophisticated multi-stage methods and runs in real-time.

Convolutional Neural Networks on non-uniform geometrical signals using Euclidean spectral transformation

2019-01-07
Chiyu Max Jiang , Dequan Wang , Jingwei Huang +2 more
Convolutional Neural Networks (CNN) have been successful in processing data signals that are uniformly sampled in the spatial domain (e.g., images). However, most data signals do not natively exist on … Convolutional Neural Networks (CNN) have been successful in processing data signals that are uniformly sampled in the spatial domain (e.g., images). However, most data signals do not natively exist on a grid, and in the process of being sampled onto a uniform physical grid suffer significant aliasing error and information loss. Moreover, signals can exist in different topological structures as, for example, points, lines, surfaces and volumes. It has been challenging to analyze signals with mixed topologies (for example, point cloud with surface mesh). To this end, we develop mathematical formulations for Non-Uniform Fourier Transforms (NUFT) to directly, and optimally, sample nonuniform data signals of different topologies defined on a simplex mesh into the spectral domain with no spatial sampling error. The spectral transform is performed in the Euclidean space, which removes the translation ambiguity from works on the graph spectrum. Our representation has four distinct advantages: (1) the process causes no spatial sampling error during the initial sampling, (2) the generality of this approach provides a unified framework for using CNNs to analyze signals of mixed topologies, (3) it allows us to leverage state-of-the-art backbone CNN architectures for effective learning without having to design a particular architecture for a particular data structure in an ad-hoc fashion, and (4) the representation allows weighted meshes where each element has a different weight (i.e., texture) indicating local properties. We achieve results on par with the state-of-the-art for the 3D shape retrieval task, and a new state-of-the-art for the point cloud to surface reconstruction task.

Convolutional Neural Networks on Non-uniform Geometrical Signals Using Euclidean Spectral Transformation

2019-01-01
Chiyu Max Jiang , Dequan Wang , Jingwei Huang +2 more
Convolutional Neural Networks (CNN) have been successful in processing data signals that are uniformly sampled in the spatial domain (e.g., images). However, most data signals do not natively exist on … Convolutional Neural Networks (CNN) have been successful in processing data signals that are uniformly sampled in the spatial domain (e.g., images). However, most data signals do not natively exist on a grid, and in the process of being sampled onto a uniform physical grid suffer significant aliasing error and information loss. Moreover, signals can exist in different topological structures as, for example, points, lines, surfaces and volumes. It has been challenging to analyze signals with mixed topologies (for example, point cloud with surface mesh). To this end, we develop mathematical formulations for Non-Uniform Fourier Transforms (NUFT) to directly, and optimally, sample nonuniform data signals of different topologies defined on a simplex mesh into the spectral domain with no spatial sampling error. The spectral transform is performed in the Euclidean space, which removes the translation ambiguity from works on the graph spectrum. Our representation has four distinct advantages: (1) the process causes no spatial sampling error during the initial sampling, (2) the generality of this approach provides a unified framework for using CNNs to analyze signals of mixed topologies, (3) it allows us to leverage state-of-the-art backbone CNN architectures for effective learning without having to design a particular architecture for a particular data structure in an ad-hoc fashion, and (4) the representation allows weighted meshes where each element has a different weight (i.e., texture) indicating local properties. We achieve results on par with the state-of-the-art for the 3D shape retrieval task, and a new state-of-the-art for the point cloud to surface reconstruction task.

Dynamic Scale Inference by Entropy Minimization

2019-01-01
Dequan Wang , Evan Shelhamer , Bruno A. Olshausen +1 more
Given the variety of the visual world there is not one true scale for recognition: objects may appear at drastically different sizes across the visual field. Rather than enumerate variations … Given the variety of the visual world there is not one true scale for recognition: objects may appear at drastically different sizes across the visual field. Rather than enumerate variations across filter channels or pyramid levels, dynamic models locally predict scale and adapt receptive fields accordingly. The degree of variation and diversity of inputs makes this a difficult task. Existing methods either learn a feedforward predictor, which is not itself totally immune to the scale variation it is meant to counter, or select scales by a fixed algorithm, which cannot learn from the given task and data. We extend dynamic scale inference from feedforward prediction to iterative optimization for further adaptivity. We propose a novel entropy minimization objective for inference and optimize over task and structure parameters to tune the model to each input. Optimization during inference improves semantic segmentation accuracy and generalizes better to extreme scale variations that cause feedforward dynamic inference to falter.

Monocular Plan View Networks for Autonomous Driving

2019-01-01
Dequan Wang , Devin Coline , Qi-Zhi Cai +2 more
Convolutions on monocular dash cam videos capture spatial invariances in the image plane but do not explicitly reason about distances and depth. We propose a simple transformation of observations into … Convolutions on monocular dash cam videos capture spatial invariances in the image plane but do not explicitly reason about distances and depth. We propose a simple transformation of observations into a bird's eye view, also known as plan view, for end-to-end control. We detect vehicles and pedestrians in the first person view and project them into an overhead plan view. This representation provides an abstraction of the environment from which a deep network can easily deduce the positions and directions of entities. Additionally, the plan view enables us to leverage advances in 3D object detection in conjunction with deep policy learning. We evaluate our monocular plan view network on the photo-realistic Grand Theft Auto V simulator. A network using both a plan view and front view causes less than half as many collisions as previous detection-based methods and an order of magnitude fewer collisions than pure pixel-based policies.
View PDF Chat

Blurring the Line Between Structure and Learning to Optimize and Adapt Receptive Fields

2019-01-01
Evan Shelhamer , Dequan Wang , Trevor Darrell
The visual world is vast and varied, but its variations divide into structured and unstructured factors. We compose free-form filters and structured Gaussian filters, optimized end-to-end, to factorize deep representations … The visual world is vast and varied, but its variations divide into structured and unstructured factors. We compose free-form filters and structured Gaussian filters, optimized end-to-end, to factorize deep representations and learn both local features and their degree of locality. Our semi-structured composition is strictly more expressive than free-form filtering, and changes in its structured parameters would require changes in free-form architecture. In effect this optimizes over receptive field size and shape, tuning locality to the data and task. Dynamic inference, in which the Gaussian structure varies with the input, adapts receptive field size to compensate for local scale variation. Optimizing receptive field size improves semantic segmentation accuracy on Cityscapes by 1-2 points for strong dilated and skip architectures and by up to 10 points for suboptimal designs. Adapting receptive fields by dynamic Gaussian structure further improves results, equaling the accuracy of free-form deformation while improving efficiency.
View PDF Chat

Convolutional Neural Networks on non-uniform geometrical signals using Euclidean spectral transformation

2019-01-01
Chiyu Jiang , Dequan Wang , Jingwei Huang +2 more
Convolutional Neural Networks (CNN) have been successful in processing data signals that are uniformly sampled in the spatial domain (e.g., images). However, most data signals do not natively exist on … Convolutional Neural Networks (CNN) have been successful in processing data signals that are uniformly sampled in the spatial domain (e.g., images). However, most data signals do not natively exist on a grid, and in the process of being sampled onto a uniform physical grid suffer significant aliasing error and information loss. Moreover, signals can exist in different topological structures as, for example, points, lines, surfaces and volumes. It has been challenging to analyze signals with mixed topologies (for example, point cloud with surface mesh). To this end, we develop mathematical formulations for Non-Uniform Fourier Transforms (NUFT) to directly, and optimally, sample nonuniform data signals of different topologies defined on a simplex mesh into the spectral domain with no spatial sampling error. The spectral transform is performed in the Euclidean space, which removes the translation ambiguity from works on the graph spectrum. Our representation has four distinct advantages: (1) the process causes no spatial sampling error during the initial sampling, (2) the generality of this approach provides a unified framework for using CNNs to analyze signals of mixed topologies, (3) it allows us to leverage state-of-the-art backbone CNN architectures for effective learning without having to design a particular architecture for a particular data structure in an ad-hoc fashion, and (4) the representation allows weighted meshes where each element has a different weight (i.e., texture) indicating local properties. We achieve results on par with the state-of-the-art for the 3D shape retrieval task, and a new state-of-the-art for the point cloud to surface reconstruction task.
View PDF Chat

Objects as Points

2019-01-01
Xingyi Zhou , Dequan Wang , Philipp Krähenbühl
Detection identifies objects as axis-aligned boxes in an image. Most successful object detectors enumerate a nearly exhaustive list of potential object locations and classify each. This is wasteful, inefficient, and … Detection identifies objects as axis-aligned boxes in an image. Most successful object detectors enumerate a nearly exhaustive list of potential object locations and classify each. This is wasteful, inefficient, and requires additional post-processing. In this paper, we take a different approach. We model an object as a single point --- the center point of its bounding box. Our detector uses keypoint estimation to find center points and regresses to all other object properties, such as size, 3D location, orientation, and even pose. Our center point based approach, CenterNet, is end-to-end differentiable, simpler, faster, and more accurate than corresponding bounding box based detectors. CenterNet achieves the best speed-accuracy trade-off on the MS COCO dataset, with 28.1% AP at 142 FPS, 37.4% AP at 52 FPS, and 45.1% AP with multi-scale testing at 1.4 FPS. We use the same approach to estimate 3D bounding box in the KITTI benchmark and human pose on the COCO keypoint dataset. Our method performs competitively with sophisticated multi-stage methods and runs in real-time.

Deep Object-Centric Policies for Autonomous Driving

2018-01-01
Dequan Wang , Coline Devin , Qi-Zhi Cai +2 more
While learning visuomotor skills in an end-to-end manner is appealing, deep neural networks are often uninterpretable and fail in surprising ways. For robotics tasks, such as autonomous driving, models that … While learning visuomotor skills in an end-to-end manner is appealing, deep neural networks are often uninterpretable and fail in surprising ways. For robotics tasks, such as autonomous driving, models that explicitly represent objects may be more robust to new scenes and provide intuitive visualizations. We describe a taxonomy of "object-centric" models which leverage both object instances and end-to-end learning. In the Grand Theft Auto V simulator, we show that object-centric models outperform object-agnostic methods in scenes with other vehicles and pedestrians, even with an imperfect detector. We also demonstrate that our architectures perform well on real-world environments by evaluating on the Berkeley DeepDrive Video dataset, where an object-centric model outperforms object-agnostic models in the low-data regimes.
View PDF Chat

Joint Monocular 3D Vehicle Detection and Tracking

2018-01-01
Hou-Ning Hu , Qi-Zhi Cai , Dequan Wang +5 more
Vehicle 3D extents and trajectories are critical cues for predicting the future location of vehicles and planning future agent ego-motion based on those predictions. In this paper, we propose a … Vehicle 3D extents and trajectories are critical cues for predicting the future location of vehicles and planning future agent ego-motion based on those predictions. In this paper, we propose a novel online framework for 3D vehicle detection and tracking from monocular videos. The framework can not only associate detections of vehicles in motion over time, but also estimate their complete 3D bounding box information from a sequence of 2D images captured on a moving platform. Our method leverages 3D box depth-ordering matching for robust instance association and utilizes 3D trajectory prediction for re-identification of occluded vehicles. We also design a motion learning module based on an LSTM for more accurate long-term motion extrapolation. Our experiments on simulation, KITTI, and Argoverse datasets show that our 3D tracking pipeline offers robust data association and tracking. On Argoverse, our image-based method is significantly better for tracking 3D vehicles within 30 meters than the LiDAR-centric baseline methods.
View PDF Chat

Iterative object and part transfer for fine-grained recognition

2017-07-01
Zhiqiang Shen , Yu–Gang Jiang , Dequan Wang +1 more
The aim of fine-grained recognition is to identify sub-ordinate categories in images like different species of birds. Existing works have confirmed that, in order to capture the subtle differences across … The aim of fine-grained recognition is to identify sub-ordinate categories in images like different species of birds. Existing works have confirmed that, in order to capture the subtle differences across the categories, automatic localization of objects and parts is critical. Most approaches for object and part localization rehed on the bottom-up pipeline, where thousands of region proposals are generated and then filtered by pre-trained object/part models. This is computationally expensive and not scalable once the number of objects/parts becomes large. In this paper, we propose a nonparametric data-driven method for object and part localization. Given an unlabeled test image, our approach transfers annotations from a few similar images retrieved in the training set. In particular, we propose an iterative transfer strategy that gradually refine the predicted bounding boxes. Based on the located objects and parts, deep convolutional features are extracted for recognition. We evaluate our approach on the widely-used CUB200-2011 dataset and a new and large dataset called Birdsnap. On both datasets, we achieve better results than many state-of-the-art approaches, including a few using oracle (manually annotated) bounding boxes in the test images.
View PDF Chat

Deep Layer Aggregation

2017-01-01
Fisher Yu , Dequan Wang , Evan Shelhamer +1 more
Visual recognition requires rich representations that span levels from low to high, scales from small to large, and resolutions from fine to coarse. Even with the depth of features in … Visual recognition requires rich representations that span levels from low to high, scales from small to large, and resolutions from fine to coarse. Even with the depth of features in a convolutional network, a layer in isolation is not enough: compounding and aggregating these representations improves inference of what and where. Architectural efforts are exploring many dimensions for network backbones, designing deeper or wider architectures, but how to best aggregate layers and blocks across a network deserves further attention. Although skip connections have been incorporated to combine layers, these connections have been "shallow" themselves, and only fuse by simple, one-step operations. We augment standard architectures with deeper aggregation to better fuse information across layers. Our deep layer aggregation structures iteratively and hierarchically merge the feature hierarchy to make networks with better accuracy and fewer parameters. Experiments across architectures and tasks show that deep layer aggregation improves recognition and resolution compared to existing branching and merging schemes. The code is at https://github.com/ucbdrive/dla.

VisDA: The Visual Domain Adaptation Challenge

2017-01-01
Xingchao Peng , Ben Usman , Neela Kaushik +3 more
We present the 2017 Visual Domain Adaptation (VisDA) dataset and challenge, a large-scale testbed for unsupervised domain adaptation across visual domains. Unsupervised domain adaptation aims to solve the real-world problem … We present the 2017 Visual Domain Adaptation (VisDA) dataset and challenge, a large-scale testbed for unsupervised domain adaptation across visual domains. Unsupervised domain adaptation aims to solve the real-world problem of domain shift, where machine learning models trained on one domain must be transferred and adapted to a novel visual domain without additional supervision. The VisDA2017 challenge is focused on the simulation-to-reality shift and has two associated tasks: image classification and image segmentation. The goal in both tracks is to first train a model on simulated, synthetic data in the source domain and then adapt it to perform well on real image data in the unlabeled test domain. Our dataset is the largest one to date for cross-domain object classification, with over 280K images across 12 categories in the combined training, validation and testing domains. The image segmentation dataset is also large-scale with over 30K images across 18 categories in the three domains. We compare VisDA to existing cross-domain adaptation datasets and provide a baseline performance analysis using various domain adaptation models that are currently popular in the field.
View PDF Chat

Iterative Object and Part Transfer for Fine-Grained Recognition

2017-01-01
Zhiqiang Shen , Yu–Gang Jiang , Dequan Wang +1 more
The aim of fine-grained recognition is to identify sub-ordinate categories in images like different species of birds. Existing works have confirmed that, in order to capture the subtle differences across … The aim of fine-grained recognition is to identify sub-ordinate categories in images like different species of birds. Existing works have confirmed that, in order to capture the subtle differences across the categories, automatic localization of objects and parts is critical. Most approaches for object and part localization relied on the bottom-up pipeline, where thousands of region proposals are generated and then filtered by pre-trained object/part models. This is computationally expensive and not scalable once the number of objects/parts becomes large. In this paper, we propose a nonparametric data-driven method for object and part localization. Given an unlabeled test image, our approach transfers annotations from a few similar images retrieved in the training set. In particular, we propose an iterative transfer strategy that gradually refine the predicted bounding boxes. Based on the located objects and parts, deep convolutional features are extracted for recognition. We evaluate our approach on the widely-used CUB200-2011 dataset and a new and large dataset called Birdsnap. On both datasets, we achieve better results than many state-of-the-art approaches, including a few using oracle (manually annotated) bounding boxes in the test images.
View PDF Chat

FCNs in the Wild: Pixel-level Adversarial and Constraint-based Adaptation

2016-01-01
Judy Hoffman , Dequan Wang , Fisher Yu +1 more
Fully convolutional models for dense prediction have proven successful for a wide range of visual tasks. Such models perform well in a supervised setting, but performance can be surprisingly poor … Fully convolutional models for dense prediction have proven successful for a wide range of visual tasks. Such models perform well in a supervised setting, but performance can be surprisingly poor under domain shifts that appear mild to a human observer. For example, training on one city and testing on another in a different geographic region and/or weather condition may result in significantly degraded performance due to pixel-level distribution shift. In this paper, we introduce the first domain adaptive semantic segmentation method, proposing an unsupervised adversarial approach to pixel prediction problems. Our method consists of both global and category specific adaptation techniques. Global domain alignment is performed using a novel semantic segmentation network with fully convolutional domain adversarial learning. This initially adapted space then enables category specific adaptation through a generalization of constrained weak learning, with explicit transfer of the spatial layout from the source to the target domains. Our approach outperforms baselines across different settings on multiple large-scale datasets, including adapting across various real city environments, different synthetic sub-domains, from simulated to real environments, and on a novel large-scale dash-cam dataset.

Modified Homotopy Analysis Method for Nonlinear Aeroelastic Behavior of Two Degree-of-Freedom Airfoils

2015-10-07
Yaobin Niu , Zhongwei Wang , Dequan Wang +1 more
In this paper, the homotopy analysis method (HAM) is extended to deal with the nonlinear aeroelastic problem of a two degree-of-freedom (DOF) airfoil. To avoid determination of the parameter for … In this paper, the homotopy analysis method (HAM) is extended to deal with the nonlinear aeroelastic problem of a two degree-of-freedom (DOF) airfoil. To avoid determination of the parameter for the complicated high-order minimization problem, a new modified HAM is proposed based on the idea of minimizing the squared residual. Using this method, the convergence-control parameter is determined by the low order squared residual of the governing equations, and then the problem is solved in a way similar to the basic HAM. The proposed method is used to solve the nonlinear aeroelastic behavior of a supersonic airfoil, with the unsteady aerodynamic load evaluated by the piston theory. Two examples are prepared, for which the frequencies and amplitudes of the limit cycles are obtained. The approximate solutions obtained are demonstrated to agree excellently the numerical solutions, meanwhile, the convergence-control parameter can be easily determined using the present approach.

Learning to Point and Count

2015-01-01
Jie Shao , Dequan Wang , Xiangyang Xue +1 more
This paper proposes the problem of point-and-count as a test case to break the what-and-where deadlock. Different from the traditional detection problem, the goal is to discover key salient points … This paper proposes the problem of point-and-count as a test case to break the what-and-where deadlock. Different from the traditional detection problem, the goal is to discover key salient points as a way to localize and count the number of objects simultaneously. We propose two alternatives, one that counts first and then point, and another that works the other way around. Fundamentally, they pivot around whether we solve "what" or "where" first. We evaluate their performance on dataset that contains multiple instances of the same class, demonstrating the potentials and their synergies. The experiences derive a few important insights that explains why this is a much harder problem than classification, including strong data bias and the inability to deal with object scales robustly in state-of-art convolutional neural networks.

Numerical Simulation of Fuel Regression Rate Behavior of Hybrid Rocket Motors

2013-01-01
Jianxin Hu , Zhi-xun Xia , Dequan Wang +1 more
The present study attempts to predict the performance of hybrid rocket motors using a physics-based comprehensive model in which all the submodels that govern the various processes that occur during … The present study attempts to predict the performance of hybrid rocket motors using a physics-based comprehensive model in which all the submodels that govern the various processes that occur during heating and combustion of solid fuel are based on physical-based mathematical models. Based on the fluid-solid coupling technique and some comprehensive physical processes during operation of hybrid rocket motors, a numerical model is developed to predict the regression rate for the solid fuel surface of the hybrid rocket motor under different operation conditions. The muti-dimensional Favre-averaged compressible turbulent Navier-Stokes equations are used as the governing equations of the reacting flow, the two-equation turbulence model is used to simulate the turbulent flow, and the eddy breakup model is used to simulate the gas combustion. The results presented are for hydroxyl-terminated-polybutadiene fuel and gaseous oxygen. The Navier-Stokes model results allow more detailed and realistic observation of the chamber flow field than is permitted by simpler boundary layer analyses. The model predications indicate that fuel surface regression rates are considerably impacted by both the size and geometry of the configuration. This study provides considerable information to the understanding of flow and combustion process in hybrid rocket motors. Methods for optimizing the hybrid combustion characteristic are put forward.
View PDF Chat
Coauthor Papers Together
Trevor Darrell 21
Evan Shelhamer 11
Qi-Zhi Cai 7
Fisher Yu 6
Philipp Krähenbühl 6
John Wawrzynek 5
Kurt Keutzer 5
Bichen Wu 5
Yaohui Cai 5
Qijing Huang 5
Yizhao Gao 5
Shaoteng Liu 5
Sayna Ebrahimi 4
Bruno A. Olshausen 4
Zhen Dong 4
Coline Devin 4
Chengzhen Meng 3
Jingwei Huang 3
Yifan Duan 3
Xiaoran Fan 3
Xiangyang Xue 3
Matthias Nießner 3
Yanyong Zhang 3
Philip Marcus 3
Shaoting Zhang 3
Joseph E. Gonzalez 3
Lianmin Zheng 2
Chenming He 2
Michael W. Mahoney 2
Judy Hoffman 2
Ji Lin 2
Yunkun Zhang 2
Min Sun 2
Hou-Ning Hu 2
Zhijie Deng 2
Jianfei Chen 2
Chiyu Max Jiang 2
Jin Gao 2
Yu‐Gang Jiang 2
Dian Chen 2
Zhiqiang Shen 2
Xingyi Zhou 2
Yifan Lu 1
Mu Zhou 1
Yukuo Cen 1
Qicheng Lao 1
Kaahan Radia 1
Siheng Chen 1
Alexandre M. Bayen 1
Kexin Ding 1

Deep Residual Learning for Image Recognition

2016-06-01
Kaiming He , Xiangyu Zhang , Shaoqing Ren +1 more
Deeper neural networks are more difficult to train. We present a residual learning framework to ease the training of networks that are substantially deeper than those used previously. We explicitly … Deeper neural networks are more difficult to train. We present a residual learning framework to ease the training of networks that are substantially deeper than those used previously. We explicitly reformulate the layers as learning residual functions with reference to the layer inputs, instead of learning unreferenced functions. We provide comprehensive empirical evidence showing that these residual networks are easier to optimize, and can gain accuracy from considerably increased depth. On the ImageNet dataset we evaluate residual nets with a depth of up to 152 layers - 8× deeper than VGG nets [40] but still having lower complexity. An ensemble of these residual nets achieves 3.57% error on the ImageNet test set. This result won the 1st place on the ILSVRC 2015 classification task. We also present analysis on CIFAR-10 with 100 and 1000 layers. The depth of representations is of central importance for many visual recognition tasks. Solely due to our extremely deep representations, we obtain a 28% relative improvement on the COCO object detection dataset. Deep residual nets are foundations of our submissions to ILSVRC & COCO 2015 competitions1, where we also won the 1st places on the tasks of ImageNet detection, ImageNet localization, COCO detection, and COCO segmentation.
View PDF Chat

ImageNet Large Scale Visual Recognition Challenge

2015-04-10
Olga Russakovsky , Jia Deng , Hao Su +7 more
View PDF Chat

Adam: A Method for Stochastic Optimization

2014-01-01
Diederik P. Kingma , Jimmy Ba
We introduce Adam, an algorithm for first-order gradient-based optimization of stochastic objective functions, based on adaptive estimates of lower-order moments. The method is straightforward to implement, is computationally efficient, has … We introduce Adam, an algorithm for first-order gradient-based optimization of stochastic objective functions, based on adaptive estimates of lower-order moments. The method is straightforward to implement, is computationally efficient, has little memory requirements, is invariant to diagonal rescaling of the gradients, and is well suited for problems that are large in terms of data and/or parameters. The method is also appropriate for non-stationary objectives and problems with very noisy and/or sparse gradients. The hyper-parameters have intuitive interpretations and typically require little tuning. Some connections to related algorithms, on which Adam was inspired, are discussed. We also analyze the theoretical convergence properties of the algorithm and provide a regret bound on the convergence rate that is comparable to the best known results under the online convex optimization framework. Empirical results demonstrate that Adam works well in practice and compares favorably to other stochastic optimization methods. Finally, we discuss AdaMax, a variant of Adam based on the infinity norm.
View PDF Chat

The Cityscapes Dataset for Semantic Urban Scene Understanding

2016-06-01
Marius Cordts , Mohamed Omran , Sebastian Ramos +6 more
Visual understanding of complex urban street scenes is an enabling factor for a wide range of applications. Object detection has benefited enormously from large-scale datasets, especially in the context of … Visual understanding of complex urban street scenes is an enabling factor for a wide range of applications. Object detection has benefited enormously from large-scale datasets, especially in the context of deep learning. For semantic urban scene understanding, however, no current dataset adequately captures the complexity of real-world urban scenes. To address this, we introduce Cityscapes, a benchmark suite and large-scale dataset to train and test approaches for pixel-level and instance-level semantic labeling. Cityscapes is comprised of a large, diverse set of stereo video sequences recorded in streets from 50 different cities. 5000 of these images have high quality pixel-level annotations, 20 000 additional images have coarse annotations to enable methods that leverage large volumes of weakly-labeled data. Crucially, our effort exceeds previous attempts in terms of dataset size, annotation richness, scene variability, and complexity. Our accompanying empirical study provides an in-depth analysis of the dataset characteristics, as well as a performance evaluation of several state-of-the-art approaches based on our benchmark.
View PDF Chat

Deformable Convolutional Networks

2017-10-01
Jifeng Dai , Haozhi Qi , Yuwen Xiong +4 more
Convolutional neural networks (CNNs) are inherently limited to model geometric transformations due to the fixed geometric structures in their building modules. In this work, we introduce two new modules to … Convolutional neural networks (CNNs) are inherently limited to model geometric transformations due to the fixed geometric structures in their building modules. In this work, we introduce two new modules to enhance the transformation modeling capability of CNNs, namely, deformable convolution and deformable RoI pooling. Both are based on the idea of augmenting the spatial sampling locations in the modules with additional offsets and learning the offsets from the target tasks, without additional supervision. The new modules can readily replace their plain counterparts in existing CNNs and can be easily trained end-to-end by standard back-propagation, giving rise to deformable convolutional networks. Extensive experiments validate the performance of our approach. For the first time, we show that learning dense spatial transformation in deep CNNs is effective for sophisticated vision tasks such as object detection and semantic segmentation. The code is released at https://github.com/msracver/Deformable-ConvNets.
View PDF Chat

You Only Look Once: Unified, Real-Time Object Detection

2016-06-01
Joseph Redmon , Santosh Divvala , Ross Girshick +1 more
We present YOLO, a new approach to object detection. Prior work on object detection repurposes classifiers to perform detection. Instead, we frame object detection as a regression problem to spatially … We present YOLO, a new approach to object detection. Prior work on object detection repurposes classifiers to perform detection. Instead, we frame object detection as a regression problem to spatially separated bounding boxes and associated class probabilities. A single neural network predicts bounding boxes and class probabilities directly from full images in one evaluation. Since the whole detection pipeline is a single network, it can be optimized end-to-end directly on detection performance. Our unified architecture is extremely fast. Our base YOLO model processes images in real-time at 45 frames per second. A smaller version of the network, Fast YOLO, processes an astounding 155 frames per second while still achieving double the mAP of other real-time detectors. Compared to state-of-the-art detection systems, YOLO makes more localization errors but is less likely to predict false positives on background. Finally, YOLO learns very general representations of objects. It outperforms other detection methods, including DPM and R-CNN, when generalizing from natural images to other domains like artwork.
View PDF Chat

Fully Convolutional Networks for Semantic Segmentation

2016-05-24
Evan Shelhamer , Jonathan Long , Trevor Darrell
Convolutional networks are powerful visual models that yield hierarchies of features. We show that convolutional networks by themselves, trained end-to-end, pixels-to-pixels, improve on the previous best result in semantic segmentation. … Convolutional networks are powerful visual models that yield hierarchies of features. We show that convolutional networks by themselves, trained end-to-end, pixels-to-pixels, improve on the previous best result in semantic segmentation. Our key insight is to build "fully convolutional" networks that take input of arbitrary size and produce correspondingly-sized output with efficient inference and learning. We define and detail the space of fully convolutional networks, explain their application to spatially dense prediction tasks, and draw connections to prior models. We adapt contemporary classification networks (AlexNet, the VGG net, and GoogLeNet) into fully convolutional networks and transfer their learned representations by fine-tuning to the segmentation task. We then define a skip architecture that combines semantic information from a deep, coarse layer with appearance information from a shallow, fine layer to produce accurate and detailed segmentations. Our fully convolutional networks achieve improved segmentation of PASCAL VOC (30% relative improvement to 67.2% mean IU on 2012), NYUDv2, SIFT Flow, and PASCAL-Context, while inference takes one tenth of a second for a typical image.
View PDF Chat

YOLO9000: Better, Faster, Stronger

2017-07-01
Joseph Redmon , Ali Farhadi
We introduce YOLO9000, a state-of-the-art, real-time object detection system that can detect over 9000 object categories. First we propose various improvements to the YOLO detection method, both novel and drawn … We introduce YOLO9000, a state-of-the-art, real-time object detection system that can detect over 9000 object categories. First we propose various improvements to the YOLO detection method, both novel and drawn from prior work. The improved model, YOLOv2, is state-of-the-art on standard detection tasks like PASCAL VOC and COCO. Using a novel, multi-scale training method the same YOLOv2 model can run at varying sizes, offering an easy tradeoff between speed and accuracy. At 67 FPS, YOLOv2 gets 76.8 mAP on VOC 2007. At 40 FPS, YOLOv2 gets 78.6 mAP, outperforming state-of-the-art methods like Faster RCNN with ResNet and SSD while still running significantly faster. Finally we propose a method to jointly train on object detection and classification. Using this method we train YOLO9000 simultaneously on the COCO detection dataset and the ImageNet classification dataset. Our joint training allows YOLO9000 to predict detections for object classes that dont have labelled detection data. We validate our approach on the ImageNet detection task. YOLO9000 gets 19.7 mAP on the ImageNet detection validation set despite only having detection data for 44 of the 200 classes. On the 156 classes not in COCO, YOLO9000 gets 16.0 mAP. YOLO9000 predicts detections for more than 9000 different object categories, all in real-time.
View PDF Chat

SSD: Single Shot MultiBox Detector

2016-01-01
Wei Liu , Dragomir Anguelov , Dumitru Erhan +4 more
View PDF Chat

Rich Feature Hierarchies for Accurate Object Detection and Semantic Segmentation

2014-06-01
Ross Girshick , Jeff Donahue , Trevor Darrell +1 more
Object detection performance, as measured on the canonical PASCAL VOC dataset, has plateaued in the last few years. The best-performing methods are complex ensemble systems that typically combine multiple low-level … Object detection performance, as measured on the canonical PASCAL VOC dataset, has plateaued in the last few years. The best-performing methods are complex ensemble systems that typically combine multiple low-level image features with high-level context. In this paper, we propose a simple and scalable detection algorithm that improves mean average precision (mAP) by more than 30% relative to the previous best result on VOC 2012 -- achieving a mAP of 53.3%. Our approach combines two key insights: (1) one can apply high-capacity convolutional neural networks (CNNs) to bottom-up region proposals in order to localize and segment objects and (2) when labeled training data is scarce, supervised pre-training for an auxiliary task, followed by domain-specific fine-tuning, yields a significant performance boost. Since we combine region proposals with CNNs, we call our method R-CNN: Regions with CNN features. We also present experiments that provide insight into what the network learns, revealing a rich hierarchy of image features. Source code for the complete system is available at http://www.cs.berkeley.edu/~rbg/rcnn.
View PDF Chat

Depth Map Prediction from a Single Image using a Multi-Scale Deep Network

2014-01-01
David Eigen , Christian Puhrsch , Rob Fergus
Predicting depth is an essential component in understanding the 3D geometry of a scene. While for stereo images local correspondence suffices for estimation, finding depth relations from a single image … Predicting depth is an essential component in understanding the 3D geometry of a scene. While for stereo images local correspondence suffices for estimation, finding depth relations from a single image is less straightforward, requiring integration of both global and local information from various cues. Moreover, the task is inherently ambiguous, with a large source of uncertainty coming from the overall scale. In this paper, we present a new method that addresses this task by employing two deep network stacks: one that makes a coarse global prediction based on the entire image, and another that refines this prediction locally. We also apply a scale-invariant error to help measure depth relations rather than scale. By leveraging the raw datasets as large sources of training data, our method achieves state-of-the-art results on both NYU Depth and KITTI, and matches detailed depth boundaries without the need for superpixelation.

Very Deep Convolutional Networks for Large-Scale Image Recognition

2014-01-01
Karen Simonyan , Andrew Zisserman
In this work we investigate the effect of the convolutional network depth on its accuracy in the large-scale image recognition setting. Our main contribution is a thorough evaluation of networks … In this work we investigate the effect of the convolutional network depth on its accuracy in the large-scale image recognition setting. Our main contribution is a thorough evaluation of networks of increasing depth using an architecture with very small (3x3) convolution filters, which shows that a significant improvement on the prior-art configurations can be achieved by pushing the depth to 16-19 weight layers. These findings were the basis of our ImageNet Challenge 2014 submission, where our team secured the first and the second places in the localisation and classification tracks respectively. We also show that our representations generalise well to other datasets, where they achieve state-of-the-art results. We have made our two best-performing ConvNet models publicly available to facilitate further research on the use of deep visual representations in computer vision.

Feature Pyramid Networks for Object Detection

2017-07-01
Tsung-Yi Lin , Piotr Dollár , Ross Girshick +3 more
Feature pyramids are a basic component in recognition systems for detecting objects at different scales. But pyramid representations have been avoided in recent object detectors that are based on deep … Feature pyramids are a basic component in recognition systems for detecting objects at different scales. But pyramid representations have been avoided in recent object detectors that are based on deep convolutional networks, partially because they are slow to compute and memory intensive. In this paper, we exploit the inherent multi-scale, pyramidal hierarchy of deep convolutional networks to construct feature pyramids with marginal extra cost. A top-down architecture with lateral connections is developed for building high-level semantic feature maps at all scales. This architecture, called a Feature Pyramid Network (FPN), shows significant improvement as a generic feature extractor in several applications. Using a basic Faster R-CNN system, our method achieves state-of-the-art single-model results on the COCO detection benchmark without bells and whistles, surpassing all existing single-model entries including those from the COCO 2016 challenge winners. In addition, our method can run at 5 FPS on a GPU and thus is a practical and accurate solution to multi-scale object detection. Code will be made publicly available.
View PDF Chat

Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift

2015-01-01
Sergey Ioffe , Christian Szegedy
Training Deep Neural Networks is complicated by the fact that the distribution of each layer's inputs changes during training, as the parameters of the previous layers change. This slows down … Training Deep Neural Networks is complicated by the fact that the distribution of each layer's inputs changes during training, as the parameters of the previous layers change. This slows down the training by requiring lower learning rates and careful parameter initialization, and makes it notoriously hard to train models with saturating nonlinearities. We refer to this phenomenon as internal covariate shift, and address the problem by normalizing layer inputs. Our method draws its strength from making normalization a part of the model architecture and performing the normalization for each training mini-batch. Batch Normalization allows us to use much higher learning rates and be less careful about initialization. It also acts as a regularizer, in some cases eliminating the need for Dropout. Applied to a state-of-the-art image classification model, Batch Normalization achieves the same accuracy with 14 times fewer training steps, and beats the original model by a significant margin. Using an ensemble of batch-normalized networks, we improve upon the best published result on ImageNet classification: reaching 4.9% top-5 validation error (and 4.8% test error), exceeding the accuracy of human raters.

Deformable ConvNets V2: More Deformable, Better Results

2019-06-01
Xizhou Zhu , Han Hu , Stephen Lin +1 more
The superior performance of Deformable Convolutional Networks arises from its ability to adapt to the geometric variations of objects. Through an examination of its adaptive behavior, we observe that while … The superior performance of Deformable Convolutional Networks arises from its ability to adapt to the geometric variations of objects. Through an examination of its adaptive behavior, we observe that while the spatial support for its neural features conforms more closely than regular ConvNets to object structure, this support may nevertheless extend well beyond the region of interest, causing features to be influenced by irrelevant image content. To address this problem, we present a reformulation of Deformable ConvNets that improves its ability to focus on pertinent image regions, through increased modeling power and stronger training. The modeling power is enhanced through a more comprehensive integration of deformable convolution within the network, and by introducing a modulation mechanism that expands the scope of deformation modeling. To effectively harness this enriched modeling capability, we guide network training via a proposed feature mimicking scheme that helps the network to learn features that reflect the object focus and classification power of R-CNN features. With the proposed contributions, this new version of Deformable ConvNets yields significant performance gains over the original model and produces leading results on the COCO benchmark for object detection and instance segmentation.
View PDF Chat

PyTorch: An Imperative Style, High-Performance Deep Learning Library

2019-01-01
Adam Paszke , Sam Gross , Francisco Massa +7 more
Deep learning frameworks have often focused on either usability or speed, but not both. PyTorch is a machine learning library that shows that these two goals are in fact compatible: … Deep learning frameworks have often focused on either usability or speed, but not both. PyTorch is a machine learning library that shows that these two goals are in fact compatible: it was designed from first principles to support an imperative and Pythonic programming style that supports code as a model, makes debugging easy and is consistent with other popular scientific computing libraries, while remaining efficient and supporting hardware accelerators such as GPUs. In this paper, we detail the principles that drove the implementation of PyTorch and how they are reflected in its architecture. We emphasize that every aspect of PyTorch is a regular Python program under the full control of its user. We also explain how the careful and pragmatic implementation of the key components of its runtime enables them to work together to achieve compelling performance. We demonstrate the efficiency of individual subsystems, as well as the overall speed of PyTorch on several commonly used benchmarks.

3D Bounding Box Estimation Using Deep Learning and Geometry

2017-07-01
Arsalan Mousavian , Dragomir Anguelov , John P. Flynn +1 more
We present a method for 3D object detection and pose estimation from a single image. In contrast to current techniques that only regress the 3D orientation of an object, our … We present a method for 3D object detection and pose estimation from a single image. In contrast to current techniques that only regress the 3D orientation of an object, our method first regresses relatively stable 3D object properties using a deep convolutional neural network and then combines these estimates with geometric constraints provided by a 2D object bounding box to produce a complete 3D bounding box. The first network output estimates the 3D object orientation using a novel hybrid discrete-continuous loss, which significantly outperforms the L2 loss. The second output regresses the 3D object dimensions, which have relatively little variance compared to alternatives and can often be predicted for many object types. These estimates, combined with the geometric constraints on translation imposed by the 2D bounding box, enable us to recover a stable and accurate 3D object pose. We evaluate our method on the challenging KITTI object detection benchmark [2] both on the official metric of 3D orientation estimation and also on the accuracy of the obtained 3D bounding boxes. Although conceptually simple, our method outperforms more complex and computationally expensive approaches that leverage semantic segmentation, instance level segmentation and flat ground priors [4] and sub-category detection [23][24]. Our discrete-continuous loss also produces state of the art results for 3D viewpoint estimation on the Pascal 3D+ dataset[26].
View PDF Chat

Deep Object-Centric Representations for Generalizable Robot Learning

2018-05-01
Coline Devin , Pieter Abbeel , Trevor Darrell +1 more
Robotic manipulation in complex open-world scenarios requires both reliable physical manipulation skills and effective and generalizable perception. In this paper, we propose using an object-centric prior and a semantic feature … Robotic manipulation in complex open-world scenarios requires both reliable physical manipulation skills and effective and generalizable perception. In this paper, we propose using an object-centric prior and a semantic feature space for the perception system of a learned policy. We devise an object-level attentional mechanism that can be used to determine relevant objects from a few trajectories or demonstrations, and then immediately incorporate those objects into a learned policy. A task-independent attention locates possible objects in the scene, and a task-specific attention identifies which objects are predictive of the trajectories. The scope of the task-specific attention is easily adjusted by showing demonstrations with distractor objects or with diverse relevant objects. Our results indicate that this approach exhibits good generalization across object instances using very few samples, and can be used to learn a variety of manipulation tasks using reinforcement learning.
View PDF Chat

Unsupervised Domain Adaptation by Backpropagation

2014-01-01
Yaroslav Ganin , Victor Lempitsky
Top-performing deep architectures are trained on massive amounts of labeled data. In the absence of labeled data for a certain task, domain adaptation often provides an attractive option given that … Top-performing deep architectures are trained on massive amounts of labeled data. In the absence of labeled data for a certain task, domain adaptation often provides an attractive option given that labeled data of similar nature but from a different domain (e.g. synthetic images) are available. Here, we propose a new approach to domain adaptation in deep architectures that can be trained on large amount of labeled data from the source domain and large amount of unlabeled data from the target domain (no labeled target-domain data is necessary). As the training progresses, the approach promotes the emergence of "deep" features that are (i) discriminative for the main learning task on the source domain and (ii) invariant with respect to the shift between the domains. We show that this adaptation behaviour can be achieved in almost any feed-forward model by augmenting it with few standard layers and a simple new gradient reversal layer. The resulting augmented architecture can be trained using standard backpropagation. Overall, the approach can be implemented with little effort using any of the deep-learning packages. The method performs very well in a series of image classification experiments, achieving adaptation effect in the presence of big domain shifts and outperforming previous state-of-the-art on Office datasets.

A Reduction of Imitation Learning and Structured Prediction to No-Regret Online Learning

2010-01-01
Stéphane Ross , Geoffrey J. Gordon , J. Andrew Bagnell
Sequential prediction problems such as imitation learning, where future observations depend on previous predictions (actions), violate the common i.i.d. assumptions made in statistical learning. This leads to poor performance in … Sequential prediction problems such as imitation learning, where future observations depend on previous predictions (actions), violate the common i.i.d. assumptions made in statistical learning. This leads to poor performance in theory and often in practice. Some recent approaches provide stronger guarantees in this setting, but remain somewhat unsatisfactory as they train either non-stationary or stochastic policies and require a large number of iterations. In this paper, we propose a new iterative algorithm, which trains a stationary deterministic policy, that can be seen as a no regret algorithm in an online learning setting. We show that any such no regret algorithm, combined with additional reduction assumptions, must find a policy with good performance under the distribution of observations it induces in such sequential settings. We demonstrate that this new approach outperforms previous approaches on two challenging imitation learning problems and a benchmark sequence labeling problem.

Interpretable Learning for Self-Driving Cars by Visualizing Causal Attention

2017-10-01
Jinkyu Kim , John Canny
Deep neural perception and control networks are likely to be a key component of self-driving vehicles. These models need to be explainable - they should provide easy-tointerpret rationales for their … Deep neural perception and control networks are likely to be a key component of self-driving vehicles. These models need to be explainable - they should provide easy-tointerpret rationales for their behavior - so that passengers, insurance companies, law enforcement, developers etc., can understand what triggered a particular behavior. Here we explore the use of visual explanations. These explanations take the form of real-time highlighted regions of an image that causally influence the network's output (steering control). Our approach is two-stage. In the first stage, we use a visual attention model to train a convolution network endto- end from images to steering angle. The attention model highlights image regions that potentially influence the network's output. Some of these are true influences, but some are spurious. We then apply a causal filtering step to determine which input regions actually influence the output. This produces more succinct visual explanations and more accurately exposes the network's behavior. We demonstrate the effectiveness of our model on three datasets totaling 16 hours of driving. We first show that training with attention does not degrade the performance of the end-to-end network. Then we show that the network causally cues on a variety of features that are used by humans while driving.
View PDF Chat

End-to-End Learning of Driving Models from Large-Scale Video Datasets

2017-07-01
Huazhe Xu , Yang Gao , Fisher Yu +1 more
Robust perception-action models should be learned from training data with diverse visual appearances and realistic behaviors, yet current approaches to deep visuomotor policy learning have been generally limited to in-situ … Robust perception-action models should be learned from training data with diverse visual appearances and realistic behaviors, yet current approaches to deep visuomotor policy learning have been generally limited to in-situ models learned from a single vehicle or simulation environment. We advocate learning a generic vehicle motion model from large scale crowd-sourced video data, and develop an end-to-end trainable architecture for learning to predict a distribution over future vehicle egomotion from instantaneous monocular camera observations and previous vehicle state. Our model incorporates a novel FCN-LSTM architecture, which can be learned from large-scale crowd-sourced vehicle action data, and leverages available scene segmentation side tasks to improve performance under a privileged learning paradigm. We provide a novel large-scale dataset of crowd-sourced driving behavior suitable for training our model, and report results predicting the driver action on held out sequences across diverse conditions.
View PDF Chat

End to End Learning for Self-Driving Cars

2016-01-01
Mariusz Bojarski , Davide Del Testa , Daniel Dworakowski +7 more
We trained a convolutional neural network (CNN) to map raw pixels from a single front-facing camera directly to steering commands. This end-to-end approach proved surprisingly powerful. With minimum training data … We trained a convolutional neural network (CNN) to map raw pixels from a single front-facing camera directly to steering commands. This end-to-end approach proved surprisingly powerful. With minimum training data from humans the system learns to drive in traffic on local roads with or without lane markings and on highways. It also operates in areas with unclear visual guidance such as in parking lots and on unpaved roads. The system automatically learns internal representations of the necessary processing steps such as detecting useful road features with only the human steering angle as the training signal. We never explicitly trained it to detect, for example, the outline of roads. Compared to explicit decomposition of the problem, such as lane marking detection, path planning, and control, our end-to-end system optimizes all processing steps simultaneously. We argue that this will eventually lead to better performance and smaller systems. Better performance will result because the internal components self-optimize to maximize overall system performance, instead of optimizing human-selected intermediate criteria, e.g., lane detection. Such criteria understandably are selected for ease of human interpretation which doesn't automatically guarantee maximum system performance. Smaller networks are possible because the system learns to solve the problem with the minimal number of processing steps. We used an NVIDIA DevBox and Torch 7 for training and an NVIDIA DRIVE(TM) PX self-driving car computer also running Torch 7 for determining where to drive. The system operates at 30 frames per second (FPS).

Deep High-Resolution Representation Learning for Visual Recognition

2019-01-01
Jingdong Wang , Ke Sun , Tianheng Cheng +7 more
High-resolution representations are essential for position-sensitive vision problems, such as human pose estimation, semantic segmentation, and object detection. Existing state-of-the-art frameworks first encode the input image as a low-resolution representation … High-resolution representations are essential for position-sensitive vision problems, such as human pose estimation, semantic segmentation, and object detection. Existing state-of-the-art frameworks first encode the input image as a low-resolution representation through a subnetwork that is formed by connecting high-to-low resolution convolutions \emph{in series} (e.g., ResNet, VGGNet), and then recover the high-resolution representation from the encoded low-resolution representation. Instead, our proposed network, named as High-Resolution Network (HRNet), maintains high-resolution representations through the whole process. There are two key characteristics: (i) Connect the high-to-low resolution convolution streams \emph{in parallel}; (ii) Repeatedly exchange the information across resolutions. The benefit is that the resulting representation is semantically richer and spatially more precise. We show the superiority of the proposed HRNet in a wide range of applications, including human pose estimation, semantic segmentation, and object detection, suggesting that the HRNet is a stronger backbone for computer vision problems. All the codes are available at~{\url{https://github.com/HRNet}}.
View PDF Chat

ShuffleNet V2: Practical Guidelines for Efficient CNN Architecture Design

2018-01-01
Ningning Ma , Xiangyu Zhang , Hai-Tao Zheng +1 more
View PDF Chat

Do We Really Need to Access the Source Data? Source Hypothesis Transfer for Unsupervised Domain Adaptation

2020-01-01
Jian Liang , Dapeng Hu , Jiashi Feng
Unsupervised domain adaptation (UDA) aims to leverage the knowledge learned from a labeled source dataset to solve similar tasks in a new unlabeled domain. Prior UDA methods typically require to … Unsupervised domain adaptation (UDA) aims to leverage the knowledge learned from a labeled source dataset to solve similar tasks in a new unlabeled domain. Prior UDA methods typically require to access the source data when learning to adapt the model, making them risky and inefficient for decentralized private data. This work tackles a practical setting where only a trained source model is available and investigates how we can effectively utilize such a model without source data to solve UDA problems. We propose a simple yet generic representation learning framework, named \emph{Source HypOthesis Transfer} (SHOT). SHOT freezes the classifier module (hypothesis) of the source model and learns the target-specific feature extraction module by exploiting both information maximization and self-supervised pseudo-labeling to implicitly align representations from the target domains to the source hypothesis. To verify its versatility, we evaluate SHOT in a variety of adaptation cases including closed-set, partial-set, and open-set domain adaptation. Experiments indicate that SHOT yields state-of-the-art results among multiple domain adaptation benchmarks.

DeepDriving: Learning Affordance for Direct Perception in Autonomous Driving

2015-12-01
Chenyi Chen , Ari Seff , Alain L. Kornhauser +1 more
Today, there are two major paradigms for vision-based autonomous driving systems: mediated perception approaches that parse an entire scene to make a driving decision, and behavior reflex approaches that directly … Today, there are two major paradigms for vision-based autonomous driving systems: mediated perception approaches that parse an entire scene to make a driving decision, and behavior reflex approaches that directly map an input image to a driving action by a regressor. In this paper, we propose a third paradigm: a direct perception approach to estimate the affordance for driving. We propose to map an input image to a small number of key perception indicators that directly relate to the affordance of a road/traffic state for driving. Our representation provides a set of compact yet complete descriptions of the scene to enable a simple controller to drive autonomously. Falling in between the two extremes of mediated perception and behavior reflex, we argue that our direct perception representation provides the right level of abstraction. To demonstrate this, we train a deep Convolutional Neural Network using recording from 12 hours of human driving in a video game and show that our model can work well to drive a car in a very diverse set of virtual environments. We also train a model for car distance estimation on the KITTI dataset. Results show that our direct perception approach can generalize well to real driving images. Source code and data are available on our project website.
View PDF Chat

Explaining How a Deep Neural Network Trained with End-to-End Learning Steers a Car

2017-04-25
Mariusz Bojarski , Philip Yeres , Anna Choromanska +4 more
As part of a complete software stack for autonomous driving, NVIDIA has created a neural-network-based system, known as PilotNet, which outputs steering angles given images of the road ahead. PilotNet … As part of a complete software stack for autonomous driving, NVIDIA has created a neural-network-based system, known as PilotNet, which outputs steering angles given images of the road ahead. PilotNet is trained using road images paired with the steering angles generated by a human driving a data-collection car. It derives the necessary domain knowledge by observing human drivers. This eliminates the need for human engineers to anticipate what is important in an image and foresee all the necessary rules for safe driving. Road tests demonstrated that PilotNet can successfully perform lane keeping in a wide variety of driving conditions, regardless of whether lane markings are present or not. The goal of the work described here is to explain what PilotNet learns and how it makes its decisions. To this end we developed a method for determining which elements in the road image most influence PilotNet's steering decision. Results show that PilotNet indeed learns to recognize relevant objects on the road. In addition to learning the obvious features such as lane markings, edges of roads, and other cars, PilotNet learns more subtle features that would be hard to anticipate and program by engineers, for example, bushes lining the edge of the road and atypical vehicle classes.

Blurring the Line Between Structure and Learning to Optimize and Adapt Receptive Fields.

2019-04-25
Evan Shelhamer , Dequan Wang , Trevor Darrell
The visual world is vast and varied, but its variations divide into structured and unstructured factors. We compose free-form filters and structured Gaussian filters, optimized end-to-end, to factorize deep representations … The visual world is vast and varied, but its variations divide into structured and unstructured factors. We compose free-form filters and structured Gaussian filters, optimized end-to-end, to factorize deep representations and learn both local features and their degree of locality. Our semi-structured composition is strictly more expressive than free-form filtering, and changes in its structured parameters would require changes in free-form architecture. In effect this optimizes over receptive field size and shape, tuning locality to the data and task. Dynamic inference, in which the Gaussian structure varies with the input, adapts receptive field size to compensate for local scale variation. Optimizing receptive field size improves semantic segmentation accuracy on Cityscapes by 1-2 points for strong dilated and skip architectures and by up to 10 points for suboptimal designs. Adapting receptive fields by dynamic Gaussian structure further improves results, equaling the accuracy of free-form deformation while improving efficiency.

Simultaneous Deep Transfer Across Domains and Tasks

2015-12-01
Eric Tzeng , Judy Hoffman , Trevor Darrell +1 more
View PDF Chat

Very Deep Convolutional Networks for Large-Scale Image Recognition

2014-09-04
Karen Simonyan , Andrew Zisserman
In this work we investigate the effect of the convolutional network depth on its accuracy in the large-scale image recognition setting. Our main contribution is a thorough evaluation of networks … In this work we investigate the effect of the convolutional network depth on its accuracy in the large-scale image recognition setting. Our main contribution is a thorough evaluation of networks of increasing depth using an architecture with very small (3x3) convolution filters, which shows that a significant improvement on the prior-art configurations can be achieved by pushing the depth to 16-19 weight layers. These findings were the basis of our ImageNet Challenge 2014 submission, where our team secured the first and the second places in the localisation and classification tracks respectively. We also show that our representations generalise well to other datasets, where they achieve state-of-the-art results. We have made our two best-performing ConvNet models publicly available to facilitate further research on the use of deep visual representations in computer vision.

Euphrates: Algorithm-SoC Co-Design for Low-Power Mobile Continuous Vision

2018-06-01
Yuhao Zhu , Anand Samajdar , Matthew Mattina +1 more
Continuous computer vision (CV) tasks increasingly rely on convolutional neural networks (CNN). However, CNNs have massive compute demands that far exceed the performance and energy constraints of mobile devices. In … Continuous computer vision (CV) tasks increasingly rely on convolutional neural networks (CNN). However, CNNs have massive compute demands that far exceed the performance and energy constraints of mobile devices. In this paper, we propose and develop an algorithm-architecture co-designed system, Euphrates, that simultaneously improves the energyefficiency and performance of continuous vision tasks. Our key observation is that changes in pixel data between consecutive frames represents visual motion. We first propose an algorithm that leverages this motion information to relax the number of expensive CNN inferences required by continuous vision applications. We co-design a mobile System-ona-Chip (SoC) architecture to maximize the efficiency of the new algorithm. The key to our architectural augmentation is to co-optimize different SoC IP blocks in the vision pipeline collectively. Specifically, we propose to expose the motion data that is naturally generated by the Image Signal Processor (ISP) early in the vision pipeline to the CNN engine. Measurement and synthesis results show that Euphrates achieves up to 66% SoC-level energy savings (4× for the vision computations), with only 1% accuracy loss.
View PDF Chat

Quantization and Training of Neural Networks for Efficient Integer-Arithmetic-Only Inference

2018-06-01
Benoit Jacob , Skirmantas Kligys , Bo Chen +5 more
The rising popularity of intelligent mobile devices and the daunting computational cost of deep learning-based models call for efficient and accurate on-device inference schemes. We propose a quantization scheme that … The rising popularity of intelligent mobile devices and the daunting computational cost of deep learning-based models call for efficient and accurate on-device inference schemes. We propose a quantization scheme that allows inference to be carried out using integer-only arithmetic, which can be implemented more efficiently than floating point inference on commonly available integer-only hardware. We also co-design a training procedure to preserve end-to-end model accuracy post quantization. As a result, the proposed quantization scheme improves the tradeoff between accuracy and on-device latency. The improvements are significant even on MobileNets, a model family known for run-time efficiency, and are demonstrated in ImageNet classification and COCO detection on popular CPUs.
View PDF Chat

Focal Loss for Dense Object Detection

2017-10-01
Tsung-Yi Lin , Priya Goyal , Ross Girshick +2 more
The highest accuracy object detectors to date are based on a two-stage approach popularized by R-CNN, where a classifier is applied to a sparse set of candidate object locations. In … The highest accuracy object detectors to date are based on a two-stage approach popularized by R-CNN, where a classifier is applied to a sparse set of candidate object locations. In contrast, one-stage detectors that are applied over a regular, dense sampling of possible object locations have the potential to be faster and simpler, but have trailed the accuracy of two-stage detectors thus far. In this paper, we investigate why this is the case. We discover that the extreme foreground-background class imbalance encountered during training of dense detectors is the central cause. We propose to address this class imbalance by reshaping the standard cross entropy loss such that it down-weights the loss assigned to well-classified examples. Our novel Focal Loss focuses training on a sparse set of hard examples and prevents the vast number of easy negatives from overwhelming the detector during training. To evaluate the effectiveness of our loss, we design and train a simple dense detector we call RetinaNet. Our results show that when trained with the focal loss, RetinaNet is able to match the speed of previous one-stage detectors while surpassing the accuracy of all existing state-of-the-art two-stage detectors.
View PDF Chat

End-to-End Driving Via Conditional Imitation Learning

2018-05-01
Felipe Codevilla , Matthias Müller , Antonio M. López +2 more
Deep networks trained on demonstrations of human driving have learned to follow roads and avoid obstacles. However, driving policies trained via imitation learning cannot be controlled at test time. A … Deep networks trained on demonstrations of human driving have learned to follow roads and avoid obstacles. However, driving policies trained via imitation learning cannot be controlled at test time. A vehicle trained end-to-end to imitate an expert cannot be guided to take a specific turn at an upcoming intersection. This limits the utility of such systems. We propose to condition imitation learning on high-level command input. At test time, the learned driving policy functions as a chauffeur that handles sensorimotor coordination but continues to respond to navigational commands. We evaluate different architectures for conditional imitation learning in vision-based driving. We conduct experiments in realistic three-dimensional simulations of urban driving and on a 1/5 scale robotic truck that is trained to drive in a residential area. Both systems drive based on visual input yet remain responsive to high-level navigational commands.
View PDF Chat

Multi-view 3D Object Detection Network for Autonomous Driving

2017-07-01
Xiaozhi Chen , Huimin Ma , Ji Wan +2 more
This paper aims at high-accuracy 3D object detection in autonomous driving scenario. We propose Multi-View 3D networks (MV3D), a sensory-fusion framework that takes both LIDAR point cloud and RGB images … This paper aims at high-accuracy 3D object detection in autonomous driving scenario. We propose Multi-View 3D networks (MV3D), a sensory-fusion framework that takes both LIDAR point cloud and RGB images as input and predicts oriented 3D bounding boxes. We encode the sparse 3D point cloud with a compact multi-view representation. The network is composed of two subnetworks: one for 3D object proposal generation and another for multi-view feature fusion. The proposal network generates 3D candidate boxes efficiently from the birds eye view representation of 3D point cloud. We design a deep fusion scheme to combine region-wise features from multiple views and enable interactions between intermediate layers of different paths. Experiments on the challenging KITTI benchmark show that our approach outperforms the state-of-the-art by around 25% and 30% AP on the tasks of 3D localization and 3D detection. In addition, for 2D detection, our approach obtains 14.9% higher AP than the state-of-the-art on the hard data among the LIDAR-based methods.
View PDF Chat

Path Aggregation Network for Instance Segmentation

2018-06-01
Shu Liu , Lu Qi , Haifang Qin +2 more
The way that information propagates in neural networks is of great importance. In this paper, we propose Path Aggregation Network (PANet) aiming at boosting information flow in proposal-based instance segmentation … The way that information propagates in neural networks is of great importance. In this paper, we propose Path Aggregation Network (PANet) aiming at boosting information flow in proposal-based instance segmentation framework. Specifically, we enhance the entire feature hierarchy with accurate localization signals in lower layers by bottom-up path augmentation, which shortens the information path between lower layers and topmost feature. We present adaptive feature pooling, which links feature grid and all feature levels to make useful information in each level propagate directly to following proposal subnetworks. A complementary branch capturing different views for each proposal is created to further improve mask prediction. These improvements are simple to implement, with subtle extra computational overhead. Yet they are useful and make our PANet reach the 1st place in the COCO 2017 Challenge Instance Segmentation task and the 2nd place in Object Detection task without large-batch training. PANet is also state-of-the-art on MVD and Cityscapes.
View PDF Chat

Learning Transferable Features with Deep Adaptation Networks

2015-02-10
Mingsheng Long , Yue Cao , Jianmin Wang +1 more
Recent studies reveal that a deep neural network can learn transferable features which generalize well to novel tasks for domain adaptation. However, as deep features eventually transition from general to … Recent studies reveal that a deep neural network can learn transferable features which generalize well to novel tasks for domain adaptation. However, as deep features eventually transition from general to specific along the network, the feature transferability drops significantly in higher layers with increasing domain discrepancy. Hence, it is important to formally reduce the dataset bias and enhance the transferability in task-specific layers. In this paper, we propose a new Deep Adaptation Network (DAN) architecture, which generalizes deep convolutional neural network to the domain adaptation scenario. In DAN, hidden representations of all task-specific layers are embedded in a reproducing kernel Hilbert space where the mean embeddings of different domain distributions can be explicitly matched. The domain discrepancy is further reduced using an optimal multi-kernel selection method for mean embedding matching. DAN can learn transferable features with statistical guarantees, and can scale linearly by unbiased estimate of kernel embedding. Extensive empirical evidence shows that the proposed architecture yields state-of-the-art image classification error rates on standard domain adaptation benchmarks.

Xception: Deep Learning with Depthwise Separable Convolutions

2017-07-01
François Chollet
We present an interpretation of Inception modules in convolutional neural networks as being an intermediate step in-between regular convolution and the depthwise separable convolution operation (a depthwise convolution followed by … We present an interpretation of Inception modules in convolutional neural networks as being an intermediate step in-between regular convolution and the depthwise separable convolution operation (a depthwise convolution followed by a pointwise convolution). In this light, a depthwise separable convolution can be understood as an Inception module with a maximally large number of towers. This observation leads us to propose a novel deep convolutional neural network architecture inspired by Inception, where Inception modules have been replaced with depthwise separable convolutions. We show that this architecture, dubbed Xception, slightly outperforms Inception V3 on the ImageNet dataset (which Inception V3 was designed for), and significantly outperforms Inception V3 on a larger image classification dataset comprising 350 million images and 17,000 classes. Since the Xception architecture has the same number of parameters as Inception V3, the performance gains are not due to increased capacity but rather to a more efficient use of model parameters.
View PDF Chat

Fully Convolutional Networks for Semantic Segmentation

2016-01-01
Jonathan Long , Evan Shelhamer , Trevor Darrell
Convolutional networks are powerful visual models that yield hierarchies of features. We show that convolutional networks by themselves, trained end-to-end, pixels-to-pixels, improve on the previous best result in semantic segmentation. … Convolutional networks are powerful visual models that yield hierarchies of features. We show that convolutional networks by themselves, trained end-to-end, pixels-to-pixels, improve on the previous best result in semantic segmentation. Our key insight is to build "fully convolutional" networks that take input of arbitrary size and produce correspondingly-sized output with efficient inference and learning. We define and detail the space of fully convolutional networks, explain their application to spatially dense prediction tasks, and draw connections to prior models. We adapt contemporary classification networks (AlexNet, the VGG net, and GoogLeNet) into fully convolutional networks and transfer their learned representations by fine-tuning to the segmentation task. We then define a skip architecture that combines semantic information from a deep, coarse layer with appearance information from a shallow, fine layer to produce accurate and detailed segmentations. Our fully convolutional network achieves improved segmentation of PASCAL VOC (30% relative improvement to 67.2% mean IU on 2012), NYUDv2, SIFT Flow, and PASCAL-Context, while inference takes one tenth of a second for a typical image.
View PDF Chat

Some Improvements on Deep Convolutional Neural Network Based Image Classification

2013-12-19
Andrew Howard
We investigate multiple techniques to improve upon the current state of the art deep convolutional neural network based image classification pipeline. The techiques include adding more image transformations to training … We investigate multiple techniques to improve upon the current state of the art deep convolutional neural network based image classification pipeline. The techiques include adding more image transformations to training data, adding more transformations to generate additional predictions at test time and using complementary models applied to higher resolution images. This paper summarizes our entry in the Imagenet Large Scale Visual Recognition Challenge 2013. Our system achieved a top 5 classification error rate of 13.55% using no external data which is over a 20% relative improvement on the previous year's winner.

Subcategory-Aware Convolutional Neural Networks for Object Proposals and Detection

2017-03-01
Yu Xiang , Wongun Choi , Yuanqing Lin +1 more
In Convolutional Neural Network (CNN)-based object detection methods, region proposal becomes a bottleneck when objects exhibit significant scale variation, occlusion or truncation. In addition, these methods mainly focus on 2D … In Convolutional Neural Network (CNN)-based object detection methods, region proposal becomes a bottleneck when objects exhibit significant scale variation, occlusion or truncation. In addition, these methods mainly focus on 2D object detection and cannot estimate detailed properties of objects. In this paper, we propose subcategory-aware CNNs for object detection. We introduce a novel region proposal network that uses subcategory information to guide the proposal generating process, and a new detection network for joint detection and subcategory classification. By using subcategories related to object pose, we achieve state of-the-art performance on both detection and pose estimation on commonly used benchmarks.
View PDF Chat

DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFs

2017-04-27
Liang-Chieh Chen , George Papandreou , Iasonas Kokkinos +2 more
In this work we address the task of semantic image segmentation with Deep Learning and make three main contributions that are experimentally shown to have substantial practical merit. First, we … In this work we address the task of semantic image segmentation with Deep Learning and make three main contributions that are experimentally shown to have substantial practical merit. First, we highlight convolution with upsampled filters, or 'atrous convolution', as a powerful tool in dense prediction tasks. Atrous convolution allows us to explicitly control the resolution at which feature responses are computed within Deep Convolutional Neural Networks. It also allows us to effectively enlarge the field of view of filters to incorporate larger context without increasing the number of parameters or the amount of computation. Second, we propose atrous spatial pyramid pooling (ASPP) to robustly segment objects at multiple scales. ASPP probes an incoming convolutional feature layer with filters at multiple sampling rates and effective fields-of-views, thus capturing objects as well as image context at multiple scales. Third, we improve the localization of object boundaries by combining methods from DCNNs and probabilistic graphical models. The commonly deployed combination of max-pooling and downsampling in DCNNs achieves invariance but has a toll on localization accuracy. We overcome this by combining the responses at the final DCNN layer with a fully connected Conditional Random Field (CRF), which is shown both qualitatively and quantitatively to improve localization performance. Our proposed "DeepLab" system sets the new state-of-art at the PASCAL VOC-2012 semantic image segmentation task, reaching 79.7 percent mIOU in the test set, and advances the results on three other datasets: PASCAL-Context, PASCAL-Person-Part, and Cityscapes. All of our code is made publicly available online.
View PDF Chat

DoReFa-Net: Training Low Bitwidth Convolutional Neural Networks with Low Bitwidth Gradients

2016-01-01
Shuchang Zhou , Yuxin Wu , Zekun Ni +3 more
We propose DoReFa-Net, a method to train convolutional neural networks that have low bitwidth weights and activations using low bitwidth parameter gradients. In particular, during backward pass, parameter gradients are … We propose DoReFa-Net, a method to train convolutional neural networks that have low bitwidth weights and activations using low bitwidth parameter gradients. In particular, during backward pass, parameter gradients are stochastically quantized to low bitwidth numbers before being propagated to convolutional layers. As convolutions during forward/backward passes can now operate on low bitwidth weights and activations/gradients respectively, DoReFa-Net can use bit convolution kernels to accelerate both training and inference. Moreover, as bit convolutions can be efficiently implemented on CPU, FPGA, ASIC and GPU, DoReFa-Net opens the way to accelerate training of low bitwidth neural network on these hardware. Our experiments on SVHN and ImageNet datasets prove that DoReFa-Net can achieve comparable prediction accuracy as 32-bit counterparts. For example, a DoReFa-Net derived from AlexNet that has 1-bit weights, 2-bit activations, can be trained from scratch using 6-bit gradients to get 46.1\% top-1 accuracy on ImageNet validation set. The DoReFa-Net AlexNet model is released publicly.

Playing for Data: Ground Truth from Computer Games

2016-01-01
Stephan R. Richter , Vibhav Vineet , Stefan Roth +1 more
View PDF Chat

Locally Scale-Invariant Convolutional Neural Networks

2014-01-01
Angjoo Kanazawa , Abhishek Sharma , David W. Jacobs
Convolutional Neural Networks (ConvNets) have shown excellent results on many visual classification tasks. With the exception of ImageNet, these datasets are carefully crafted such that objects are well-aligned at similar … Convolutional Neural Networks (ConvNets) have shown excellent results on many visual classification tasks. With the exception of ImageNet, these datasets are carefully crafted such that objects are well-aligned at similar scales. Naturally, the feature learning problem gets more challenging as the amount of variation in the data increases, as the models have to learn to be invariant to certain changes in appearance. Recent results on the ImageNet dataset show that given enough data, ConvNets can learn such invariances producing very discriminative features [1]. But could we do more: use less parameters, less data, learn more discriminative features, if certain invariances were built into the learning process? In this paper we present a simple model that allows ConvNets to learn features in a locally scale-invariant manner without increasing the number of model parameters. We show on a modified MNIST dataset that when faced with scale variation, building in scale-invariance allows ConvNets to learn more discriminative features with reduced chances of over-fitting.
View PDF Chat

MMDetection: Open MMLab Detection Toolbox and Benchmark.

2019-06-17
Kai Chen , Jiaqi Wang , Jiangmiao Pang +7 more
We present MMDetection, an object detection toolbox that contains a rich set of object detection and instance segmentation methods as well as related components and modules. The toolbox started from … We present MMDetection, an object detection toolbox that contains a rich set of object detection and instance segmentation methods as well as related components and modules. The toolbox started from a codebase of MMDet team who won the detection track of COCO Challenge 2018. It gradually evolves into a unified platform that covers many popular detection methods and contemporary modules. It not only includes training and inference codes, but also provides weights for more than 200 network models. We believe this toolbox is by far the most complete detection toolbox. In this paper, we introduce the various features of this toolbox. In addition, we also conduct a benchmarking study on different methods, components, and their hyper-parameters. We wish that the toolbox and benchmark could serve the growing research community by providing a flexible toolkit to reimplement existing methods and develop their own new detectors. Code and models are available at this https URL. The project is under active development and we will keep this document updated.

The Cityscapes Dataset for Semantic Urban Scene Understanding

2016-04-06
Marius Cordts , Mohamed Omran , Sebastian Ramos +6 more
Visual understanding of complex urban street scenes is an enabling factor for a wide range of applications. Object detection has benefited enormously from large-scale datasets, especially in the context of … Visual understanding of complex urban street scenes is an enabling factor for a wide range of applications. Object detection has benefited enormously from large-scale datasets, especially in the context of deep learning. For semantic urban scene understanding, however, no current dataset adequately captures the complexity of real-world urban scenes. To address this, we introduce Cityscapes, a benchmark suite and large-scale dataset to train and test approaches for pixel-level and instance-level semantic labeling. Cityscapes is comprised of a large, diverse set of stereo video sequences recorded in streets from 50 different cities. 5000 of these images have high quality pixel-level annotations; 20000 additional images have coarse annotations to enable methods that leverage large volumes of weakly-labeled data. Crucially, our effort exceeds previous attempts in terms of dataset size, annotation richness, scene variability, and complexity. Our accompanying empirical study provides an in-depth analysis of the dataset characteristics, as well as a performance evaluation of several state-of-the-art approaches based on our benchmark.

Driving Policy Transfer via Modularity and Abstraction

2018-01-01
Matthias Müller , Alexey Dosovitskiy , Bernard Ghanem +1 more
End-to-end approaches to autonomous driving have high sample complexity and are difficult to scale to realistic urban driving. Simulation can help end-to-end driving systems by providing a cheap, safe, and … End-to-end approaches to autonomous driving have high sample complexity and are difficult to scale to realistic urban driving. Simulation can help end-to-end driving systems by providing a cheap, safe, and diverse training environment. Yet training driving policies in simulation brings up the problem of transferring such policies to the real world. We present an approach to transferring driving policies from simulation to reality via modularity and abstraction. Our approach is inspired by classic driving systems and aims to combine the benefits of modular architectures and end-to-end deep learning approaches. The key idea is to encapsulate the driving policy such that it is not directly exposed to raw perceptual input or low-level vehicle dynamics. We evaluate the presented approach in simulated urban environments and in the real world. In particular, we transfer a driving policy trained in simulation to a 1/5-scale robotic truck that is deployed in a variety of conditions, with no finetuning, on two continents. The supplementary video can be viewed at https://youtu.be/BrMDJqI6H5U

End-to-end Learning of Multi-sensor 3D Tracking by Detection

2018-05-01
Davi Frossard , Raquel Urtasun
In this paper we propose a novel approach to tracking by detection that can exploit both cameras as well as LIDAR data to produce very accurate 3D trajectories. Towards this … In this paper we propose a novel approach to tracking by detection that can exploit both cameras as well as LIDAR data to produce very accurate 3D trajectories. Towards this goal, we formulate the problem as a linear program that can be solved exactly, and learn convolutional networks for detection as well as matching in an end-to-end manner. We evaluate our model in the challenging KITTI dataset and show very competitive results.
View PDF Chat