Search

Non-exemplar class-incremental learning (NECIL) is to resist catastrophic forgetting without saving old class samples. Prior methodologies generally employ simple rules to generate features for replaying, suffering from large distribution gap between replayed features and real ones. To address the aforementioned issue, we propose a simple, yet effective \textbf{Diff}usion-based \textbf{F}eature \textbf{R}eplay (\textbf{DiffFR}) method for NECIL. First, to alleviate the limited representational capacity caused by fixing the feature extractor, we employ Siamese-based self-supervised learning for initial generalizable features. Second, we devise diffusion models to generate class-representative features highly similar to real features, which provides an effective way for exemplar-free knowledge memorization. Third, we introduce prototype calibration to direct the diffusion model's focus towards learning the distribution shapes of features, rather than the entire distribution. Extensive experiments on public datasets demonstrate significant performance gains of our DiffFR, outperforming the state-of-the-art NECIL methods by 3.0\% in average. The code will be made publicly available soon.

Incremental object detection (IOD) aims to sequentially learn new classes, while maintaining the capability to locate and identify old ones. As the training data arrives with annotations only with new classes, IOD suffers from catastrophic forgetting. Prior methodologies mainly tackle the forgetting issue through knowledge distillation and exemplar replay, ignoring the conflict between limited model capacity and increasing knowledge. In this paper, we explore \textit{dynamic object queries} for incremental object detection built on Transformer architecture. We propose the \textbf{Dy}namic object \textbf{Q}uery-based \textbf{DE}tection \textbf{TR}ansformer (DyQ-DETR), which incrementally expands the model representation ability to achieve stability-plasticity tradeoff. First, a new set of learnable object queries are fed into the decoder to represent new classes. These new object queries are aggregated with those from previous phases to adapt both old and new knowledge well. Second, we propose the isolated bipartite matching for object queries in different phases, based on disentangled self-attention. The interaction among the object queries at different phases is eliminated to reduce inter-class confusion. Thanks to the separate supervision and computation over object queries, we further present the risk-balanced partial calibration for effective exemplar replay. Extensive experiments demonstrate that DyQ-DETR significantly surpasses the state-of-the-art methods, with limited parameter overhead. Code will be made publicly available.

Subwavelength photonic structures and metamaterials provide revolutionary approaches for controlling light. The inverse design methods proposed for these subwavelength structures are vital to the development of new photonic devices. However, most of the existing inverse design methods cannot realize direct mapping from optical properties to photonic structures but instead rely on forward simulation methods to perform iterative optimization. In this work, we exploit the powerful generative abilities of artificial intelligence (AI) and propose a practical inverse design method based on latent diffusion models. Our method maps directly the optical properties to structures without the requirement of forward simulation and iterative optimization. Here, the given optical properties can work as "prompts" and guide the constructed model to correctly "draw" the required photonic structures. Experiments show that our direct mapping-based inverse design method can generate subwavelength photonic structures at high fidelity while following the given optical properties. This may change the method used for optical design and greatly accelerate the research on new photonic devices.

In the current state of 3D object detection research, the severe scarcity of annotated 3D data, substantial disparities across different data modalities, and the absence of a unified architecture, have impeded the progress towards the goal of universality. In this paper, we propose \textbf{OV-Uni3DETR}, a unified open-vocabulary 3D detector via cycle-modality propagation. Compared with existing 3D detectors, OV-Uni3DETR offers distinct advantages: 1) Open-vocabulary 3D detection: During training, it leverages various accessible data, especially extensive 2D detection images, to boost training diversity. During inference, it can detect both seen and unseen classes. 2) Modality unifying: It seamlessly accommodates input data from any given modality, effectively addressing scenarios involving disparate modalities or missing sensor information, thereby supporting test-time modality switching. 3) Scene unifying: It provides a unified multi-modal model architecture for diverse scenes collected by distinct sensors. Specifically, we propose the cycle-modality propagation, aimed at propagating knowledge bridging 2D and 3D modalities, to support the aforementioned functionalities. 2D semantic knowledge from large-vocabulary learning guides novel class discovery in the 3D domain, and 3D geometric knowledge provides localization supervision for 2D detection images. OV-Uni3DETR achieves the state-of-the-art performance on various scenarios, surpassing existing methods by more than 6\% on average. Its performance using only RGB images is on par with or even surpasses that of previous point cloud based methods. Code and pre-trained models will be released later., Comment: Accepted by ECCV2024

Numerous point-cloud understanding techniques focus on whole entities and have succeeded in obtaining satisfactory results and limited sparsity tolerance. However, these methods are generally sensitive to incomplete point clouds that are scanned with flaws or large gaps. To address this issue, in this paper, we propose an end-to-end architecture that compensates for and identifies partial point clouds on the fly. First, we propose a cascaded solution that integrates both the upstream and downstream networks simultaneously, allowing the task-oriented downstream to identify the points generated by the completion-oriented upstream. These two streams complement each other, resulting in improved performance for both completion and downstream-dependent tasks. Second, to explicitly understand the predicted points' pattern, we introduce hierarchical self-distillation (HSD), which can be applied to arbitrary hierarchy-based point cloud methods. HSD ensures that the deepest classifier with a larger perceptual field and longer code length provides additional regularization to intermediate ones rather than simply aggregating the multi-scale features, and therefore maximizing the mutual information between a teacher and students. We show the advantage of the self-distillation process in the hyperspaces based on the information bottleneck principle. On the classification task, our proposed method performs competitively on the synthetic dataset and achieves superior results on the challenging real-world benchmark when compared to the state-of-the-art models. Additional experiments also demonstrate the superior performance and generality of our framework on the part segmentation task., Comment: Currently under review. Previously submitted to AAAI and got frustrated. Decisions: 1x weak reject, 2x weak accept, and 1 accept

Existing point cloud based 3D detectors are designed for the particular scene, either indoor or outdoor ones. Because of the substantial differences in object distribution and point density within point clouds collected from various environments, coupled with the intricate nature of 3D metrics, there is still a lack of a unified network architecture that can accommodate diverse scenes. In this paper, we propose Uni3DETR, a unified 3D detector that addresses indoor and outdoor 3D detection within the same framework. Specifically, we employ the detection transformer with point-voxel interaction for object prediction, which leverages voxel features and points for cross-attention and behaves resistant to the discrepancies from data. We then propose the mixture of query points, which sufficiently exploits global information for dense small-range indoor scenes and local information for large-range sparse outdoor ones. Furthermore, our proposed decoupled IoU provides an easy-to-optimize training target for localization by disentangling the xy and z space. Extensive experiments validate that Uni3DETR exhibits excellent performance consistently on both indoor and outdoor 3D detection. In contrast to previous specialized detectors, which may perform well on some particular datasets but suffer a substantial degradation on different scenes, Uni3DETR demonstrates the strong generalization ability under heterogeneous conditions (Fig. 1). Codes are available at \href{https://github.com/zhenyuw16/Uni3DETR}{https://github.com/zhenyuw16/Uni3DETR}., Comment: Accepted by NeurIPS 2023

For object re-identification (re-ID), learning from synthetic data has become a promising strategy to cheaply acquire large-scale annotated datasets and effective models, with few privacy concerns. Many interesting research problems arise from this strategy, e.g., how to reduce the domain gap between synthetic source and real-world target. To facilitate developing more new approaches in learning from synthetic data, we introduce the Alice benchmarks, large-scale datasets providing benchmarks as well as evaluation protocols to the research community. Within the Alice benchmarks, two object re-ID tasks are offered: person and vehicle re-ID. We collected and annotated two challenging real-world target datasets: AlicePerson and AliceVehicle, captured under various illuminations, image resolutions, etc. As an important feature of our real target, the clusterability of its training set is not manually guaranteed to make it closer to a real domain adaptation test scenario. Correspondingly, we reuse existing PersonX and VehicleX as synthetic source domains. The primary goal is to train models from synthetic data that can work effectively in the real world. In this paper, we detail the settings of Alice benchmarks, provide an analysis of existing commonly-used domain adaptation methods, and discuss some interesting future directions. An online server has been set up for the community to evaluate methods conveniently and fairly. Datasets and the online server details are available at https://sites.google.com/view/alice-benchmarks., Comment: ICLR 2024. Datasets and the online server details are available at https://sites.google.com/view/alice-benchmarks

This paper aims to learn a domain-generalizable (DG) person re-identification (ReID) representation from large-scale videos \textbf{without any annotation}. Prior DG ReID methods employ limited labeled data for training due to the high cost of annotation, which restricts further advances. To overcome the barriers of data and annotation, we propose to utilize large-scale unsupervised data for training. The key issue lies in how to mine identity information. To this end, we propose an Identity-seeking Self-supervised Representation learning (ISR) method. ISR constructs positive pairs from inter-frame images by modeling the instance association as a maximum-weight bipartite matching problem. A reliability-guided contrastive loss is further presented to suppress the adverse impact of noisy positive pairs, ensuring that reliable positive pairs dominate the learning process. The training cost of ISR scales approximately linearly with the data size, making it feasible to utilize large-scale data for training. The learned representation exhibits superior generalization ability. \textbf{Without human annotation and fine-tuning, ISR achieves 87.0\% Rank-1 on Market-1501 and 56.4\% Rank-1 on MSMT17}, outperforming the best supervised domain-generalizable method by 5.0\% and 19.5\%, respectively. In the pre-training$\rightarrow$fine-tuning scenario, ISR achieves state-of-the-art performance, with 88.4\% Rank-1 on MSMT17. The code is at \url{https://github.com/dcp15/ISR_ICCV2023_Oral}., Comment: ICCV 2023 Oral

Robotic grasping faces new challenges in human-robot-interaction scenarios. We consider the task that the robot grasps a target object designated by human's language directives. The robot not only needs to locate a target based on vision-and-language information, but also needs to predict the reasonable grasp pose candidate at various views and postures. In this work, we propose a novel interactive grasp policy, named Visual-Lingual-Grasp (VL-Grasp), to grasp the target specified by human language. First, we build a new challenging visual grounding dataset to provide functional training data for robotic interactive perception in indoor environments. Second, we propose a 6-Dof interactive grasp policy combined with visual grounding and 6-Dof grasp pose detection to extend the universality of interactive grasping. Third, we design a grasp pose filter module to enhance the performance of the policy. Experiments demonstrate the effectiveness and extendibility of the VL-Grasp in real world. The VL-Grasp achieves a success rate of 72.5\% in different indoor scenes. The code and dataset is available at https://github.com/luyh20/VL-Grasp., Comment: 8 pages, 4 figures, IROS 2023

Video anomaly detection under weak supervision presents significant challenges, particularly due to the lack of frame-level annotations during training. While prior research has utilized graph convolution networks and self-attention mechanisms alongside multiple instance learning (MIL)-based classification loss to model temporal relations and learn discriminative features, these methods often employ multi-branch architectures to capture local and global dependencies separately, resulting in increased parameters and computational costs. Moreover, the coarse-grained interclass separability provided by the binary constraint of MIL-based loss neglects the fine-grained discriminability within anomalous classes. In response, this paper introduces a weakly supervised anomaly detection framework that focuses on efficient context modeling and enhanced semantic discriminability. We present a Temporal Context Aggregation (TCA) module that captures comprehensive contextual information by reusing the similarity matrix and implementing adaptive fusion. Additionally, we propose a Prompt-Enhanced Learning (PEL) module that integrates semantic priors using knowledge-based prompts to boost the discriminative capacity of context features while ensuring separability between anomaly sub-classes. Extensive experiments validate the effectiveness of our method's components, demonstrating competitive performance with reduced parameters and computational effort on three challenging benchmarks: UCF-Crime, XD-Violence, and ShanghaiTech datasets. Notably, our approach significantly improves the detection accuracy of certain anomaly sub-classes, underscoring its practical value and efficacy. Our code is available at: https://github.com/yujiangpu20/PEL4VAD., Comment: 13 pages, 9 figures

Convolutional neural networks (CNNs) are representative models of artificial neural networks (ANNs), that form the backbone of modern computer vision. However, the considerable power consumption and limited computing speed of electrical computing platforms restrict further development of CNNs. Optical neural networks are considered the next-generation physical implementations of ANNs to break the bottleneck. This study proposes a spectral convolutional neural network (SCNN) with the function of matter meta-imaging, namely identifying the composition of matter and mapping its distribution in space. This SCNN includes an optical convolutional layer (OCL) and a reconfigurable electrical backend. The OCL is implemented by integrating very large-scale, pixel-aligned metasurfaces on a CMOS image sensor, which accepts 3D raw datacubes of natural images, containing two-spatial and one-spectral dimensions, at megapixels directly as input to realize the matter meta-imaging. This unique optoelectronic framework empowers in-sensor optical analog computing at extremely high energy efficiency eliminating the need for coherent light sources and greatly reducing the computing load of the electrical backend. We employed the SCNN framework on several real-world complex tasks. It achieved accuracies of 96.4% and 100% for pathological diagnosis and real-time face anti-spoofing at video rate, respectively. The SCNN framework, with an unprecedented new function of substance identification, provides a feasible optoelectronic and integrated optical CNN implementation for edge devices or cellphones with limited computing capabilities, facilitating diverse applications, such as intelligent robotics, industrial automation, medical diagnosis, and astronomy.

In this paper, we formally address universal object detection, which aims to detect every scene and predict every category. The dependence on human annotations, the limited visual information, and the novel categories in the open world severely restrict the universality of traditional detectors. We propose UniDetector, a universal object detector that has the ability to recognize enormous categories in the open world. The critical points for the universality of UniDetector are: 1) it leverages images of multiple sources and heterogeneous label spaces for training through the alignment of image and text spaces, which guarantees sufficient information for universal representations. 2) it generalizes to the open world easily while keeping the balance between seen and unseen classes, thanks to abundant information from both vision and language modalities. 3) it further promotes the generalization ability to novel categories through our proposed decoupling training manner and probability calibration. These contributions allow UniDetector to detect over 7k categories, the largest measurable category size so far, with only about 500 classes participating in training. Our UniDetector behaves the strong zero-shot generalization ability on large-vocabulary datasets like LVIS, ImageNetBoxes, and VisualGenome - it surpasses the traditional supervised baselines by more than 4\% on average without seeing any corresponding images. On 13 public detection datasets with various scenes, UniDetector also achieves state-of-the-art performance with only a 3\% amount of training data., Comment: Accepted by CVPR2023

In this paper, we are interested in learning a generalizable person re-identification (re-ID) representation from unlabeled videos. Compared with 1) the popular unsupervised re-ID setting where the training and test sets are typically under the same domain, and 2) the popular domain generalization (DG) re-ID setting where the training samples are labeled, our novel scenario combines their key challenges: the training samples are unlabeled, and collected form various domains which do no align with the test domain. In other words, we aim to learn a representation in an unsupervised manner and directly use the learned representation for re-ID in novel domains. To fulfill this goal, we make two main contributions: First, we propose Cycle Association (CycAs), a scalable self-supervised learning method for re-ID with low training complexity; and second, we construct a large-scale unlabeled re-ID dataset named LMP-video, tailored for the proposed method. Specifically, CycAs learns re-ID features by enforcing cycle consistency of instance association between temporally successive video frame pairs, and the training cost is merely linear to the data size, making large-scale training possible. On the other hand, the LMP-video dataset is extremely large, containing 50 million unlabeled person images cropped from over 10K Youtube videos, therefore is sufficient to serve as fertile soil for self-supervised learning. Trained on LMP-video, we show that CycAs learns good generalization towards novel domains. The achieved results sometimes even outperform supervised domain generalizable models. Remarkably, CycAs achieves 82.2% Rank-1 on Market-1501 and 49.0% Rank-1 on MSMT17 with zero human annotation, surpassing state-of-the-art supervised DG re-ID methods. Moreover, we also demonstrate the superiority of CycAs under the canonical unsupervised re-ID and the pretrain-and-finetune scenarios.

Current person image retrieval methods have achieved great improvements in accuracy metrics. However, they rarely describe the reliability of the prediction. In this paper, we propose an Uncertainty-Aware Learning (UAL) method to remedy this issue. UAL aims at providing reliability-aware predictions by considering data uncertainty and model uncertainty simultaneously. Data uncertainty captures the ``noise" inherent in the sample, while model uncertainty depicts the model's confidence in the sample's prediction. Specifically, in UAL, (1) we propose a sampling-free data uncertainty learning method to adaptively assign weights to different samples during training, down-weighting the low-quality ambiguous samples. (2) we leverage the Bayesian framework to model the model uncertainty by assuming the parameters of the network follow a Bernoulli distribution. (3) the data uncertainty and the model uncertainty are jointly learned in a unified network, and they serve as two fundamental criteria for the reliability assessment: if a probe is high-quality (low data uncertainty) and the model is confident in the prediction of the probe (low model uncertainty), the final ranking will be assessed as reliable. Experiments under the risk-controlled settings and the multi-query settings show the proposed reliability assessment is effective. Our method also shows superior performance on three challenging benchmarks under the vanilla single query settings.

A mainstream type of current self-supervised learning methods pursues a general-purpose representation that can be well transferred to downstream tasks, typically by optimizing on a given pretext task such as instance discrimination. In this work, we argue that existing pretext tasks inevitably introduce biases into the learned representation, which in turn leads to biased transfer performance on various downstream tasks. To cope with this issue, we propose Maximum Entropy Coding (MEC), a more principled objective that explicitly optimizes on the structure of the representation, so that the learned representation is less biased and thus generalizes better to unseen downstream tasks. Inspired by the principle of maximum entropy in information theory, we hypothesize that a generalizable representation should be the one that admits the maximum entropy among all plausible representations. To make the objective end-to-end trainable, we propose to leverage the minimal coding length in lossy data coding as a computationally tractable surrogate for the entropy, and further derive a scalable reformulation of the objective that allows fast computation. Extensive experiments demonstrate that MEC learns a more generalizable representation than previous methods based on specific pretext tasks. It achieves state-of-the-art performance consistently on various downstream tasks, including not only ImageNet linear probe, but also semi-supervised classification, object detection, instance segmentation, and object tracking. Interestingly, we show that existing batch-wise and feature-wise self-supervised objectives could be seen equivalent to low-order approximations of MEC. Code and pre-trained models are available at https://github.com/xinliu20/MEC., Comment: Accepted by NeurIPS 2022

Image guided depth completion aims to recover per-pixel dense depth maps from sparse depth measurements with the help of aligned color images, which has a wide range of applications from robotics to autonomous driving. However, the 3D nature of sparse-to-dense depth completion has not been fully explored by previous methods. In this work, we propose a Graph Convolution based Spatial Propagation Network (GraphCSPN) as a general approach for depth completion. First, unlike previous methods, we leverage convolution neural networks as well as graph neural networks in a complementary way for geometric representation learning. In addition, the proposed networks explicitly incorporate learnable geometric constraints to regularize the propagation process performed in three-dimensional space rather than in two-dimensional plane. Furthermore, we construct the graph utilizing sequences of feature patches, and update it dynamically with an edge attention module during propagation, so as to better capture both the local neighboring features and global relationships over long distance. Extensive experiments on both indoor NYU-Depth-v2 and outdoor KITTI datasets demonstrate that our method achieves the state-of-the-art performance, especially when compared in the case of using only a few propagation steps. Code and models are available at the project page., Comment: Accepted by ECCV 2022

We propose a task we name Portrait Interpretation and construct a dataset named Portrait250K for it. Current researches on portraits such as human attribute recognition and person re-identification have achieved many successes, but generally, they: 1) may lack mining the interrelationship between various tasks and the possible benefits it may bring; 2) design deep models specifically for each task, which is inefficient; 3) may be unable to cope with the needs of a unified model and comprehensive perception in actual scenes. In this paper, the proposed portrait interpretation recognizes the perception of humans from a new systematic perspective. We divide the perception of portraits into three aspects, namely Appearance, Posture, and Emotion, and design corresponding sub-tasks for each aspect. Based on the framework of multi-task learning, portrait interpretation requires a comprehensive description of static attributes and dynamic states of portraits. To invigorate research on this new task, we construct a new dataset that contains 250,000 images labeled with identity, gender, age, physique, height, expression, and posture of the whole body and arms. Our dataset is collected from 51 movies, hence covering extensive diversity. Furthermore, we focus on representation learning for portrait interpretation and propose a baseline that reflects our systematic perspective. We also propose an appropriate metric for this task. Our experimental results demonstrate that combining the tasks related to portrait interpretation can yield benefits. Code and dataset will be made public.

6-DoF grasp pose detection of multi-grasp and multi-object is a challenge task in the field of intelligent robot. To imitate human reasoning ability for grasping objects, data driven methods are widely studied. With the introduction of large-scale datasets, we discover that a single physical metric usually generates several discrete levels of grasp confidence scores, which cannot finely distinguish millions of grasp poses and leads to inaccurate prediction results. In this paper, we propose a hybrid physical metric to solve this evaluation insufficiency. First, we define a novel metric is based on the force-closure metric, supplemented by the measurement of the object flatness, gravity and collision. Second, we leverage this hybrid physical metric to generate elaborate confidence scores. Third, to learn the new confidence scores effectively, we design a multi-resolution network called Flatness Gravity Collision GraspNet (FGC-GraspNet). FGC-GraspNet proposes a multi-resolution features learning architecture for multiple tasks and introduces a new joint loss function that enhances the average precision of the grasp detection. The network evaluation and adequate real robot experiments demonstrate the effectiveness of our hybrid physical metric and FGC-GraspNet. Our method achieves 90.5\% success rate in real-world cluttered scenes. Our code is available at https://github.com/luyh20/FGC-GraspNet., Comment: 7 pages, 7 figures, accepted by ICRA 2022

In the paradigm of object detection, the decision head is an important part, which affects detection performance significantly. Yet how to design a high-performance decision head remains to be an open issue. In this paper, we propose a novel approach to combine decision trees and deep neural networks in an end-to-end learning manner for object detection. First, we disentangle the decision choices and prediction values by plugging soft decision trees into neural networks. To facilitate effective learning, we propose randomized decision routing with node selective and associative losses, which can boost the feature representative learning and network decision simultaneously. Second, we develop the decision head for object detection with narrow branches to generate the routing probabilities and masks, for the purpose of obtaining divergent decisions from different nodes. We name this approach as the randomized decision routing for object detection, abbreviated as R(Det)$^2$. Experiments on MS-COCO dataset demonstrate that R(Det)$^2$ is effective to improve the detection performance. Equipped with existing detectors, it achieves $1.4\sim 3.6$\% AP improvement., Comment: 10 pages, 5 figures; Accepted by CVPR2022

Current instance segmentation methods rely heavily on pixel-level annotated images. The huge cost to obtain such fully-annotated images restricts the dataset scale and limits the performance. In this paper, we formally address semi-supervised instance segmentation, where unlabeled images are employed to boost the performance. We construct a framework for semi-supervised instance segmentation by assigning pixel-level pseudo labels. Under this framework, we point out that noisy boundaries associated with pseudo labels are double-edged. We propose to exploit and resist them in a unified manner simultaneously: 1) To combat the negative effects of noisy boundaries, we propose a noise-tolerant mask head by leveraging low-resolution features. 2) To enhance the positive impacts, we introduce a boundary-preserving map for learning detailed information within boundary-relevant regions. We evaluate our approach by extensive experiments. It behaves extraordinarily, outperforming the supervised baseline by a large margin, more than 6% on Cityscapes, 7% on COCO and 4.5% on BDD100k. On Cityscapes, our method achieves comparable performance by utilizing only 30% labeled images., Comment: Accepted by CVPR2022

Clustering-based unsupervised domain adaptive (UDA) person re-identification (ReID) reduces exhaustive annotations. However, owing to unsatisfactory feature embedding and imperfect clustering, pseudo labels for target domain data inherently contain an unknown proportion of wrong ones, which would mislead feature learning. In this paper, we propose an approach named probabilistic uncertainty guided progressive label refinery (P$^2$LR) for domain adaptive person re-identification. First, we propose to model the labeling uncertainty with the probabilistic distance along with ideal single-peak distributions. A quantitative criterion is established to measure the uncertainty of pseudo labels and facilitate the network training. Second, we explore a progressive strategy for refining pseudo labels. With the uncertainty-guided alternative optimization, we balance between the exploration of target domain data and the negative effects of noisy labeling. On top of a strong baseline, we obtain significant improvements and achieve the state-of-the-art performance on four UDA ReID benchmarks. Specifically, our method outperforms the baseline by 6.5% mAP on the Duke2Market task, while surpassing the state-of-the-art method by 2.5% mAP on the Market2MSMT task., Comment: Accepted by AAAI2022

Data associations in multi-target multi-camera tracking (MTMCT) usually estimate affinity directly from re-identification (re-ID) feature distances. However, we argue that it might not be the best choice given the difference in matching scopes between re-ID and MTMCT problems. Re-ID systems focus on global matching, which retrieves targets from all cameras and all times. In contrast, data association in tracking is a local matching problem, since its candidates only come from neighboring locations and time frames. In this paper, we design experiments to verify such misfit between global re-ID feature distances and local matching in tracking, and propose a simple yet effective approach to adapt affinity estimations to corresponding matching scopes in MTMCT. Instead of trying to deal with all appearance changes, we tailor the affinity metric to specialize in ones that might emerge during data associations. To this end, we introduce a new data sampling scheme with temporal windows originally used for data associations in tracking. Minimizing the mismatch, the adaptive affinity module brings significant improvements over global re-ID distance, and produces competitive performance on CityFlow and DukeMTMC datasets., Comment: This paper appears in: IEEE Transactions on Image Processing

Semi-supervised learning aims to leverage a large amount of unlabeled data for performance boosting. Existing works primarily focus on image classification. In this paper, we delve into semi-supervised learning for object detection, where labeled data are more labor-intensive to collect. Current methods are easily distracted by noisy regions generated by pseudo labels. To combat the noisy labeling, we propose noise-resistant semi-supervised learning by quantifying the region uncertainty. We first investigate the adverse effects brought by different forms of noise associated with pseudo labels. Then we propose to quantify the uncertainty of regions by identifying the noise-resistant properties of regions over different strengths. By importing the region uncertainty quantification and promoting multipeak probability distribution output, we introduce uncertainty into training and further achieve noise-resistant learning. Experiments on both PASCAL VOC and MS COCO demonstrate the extraordinary performance of our method., Comment: Accepted by NeurIPS 2021

In this paper, we consider the problem of reference-based video super-resolution(RefVSR), i.e., how to utilize a high-resolution (HR) reference frame to super-resolve a low-resolution (LR) video sequence. The existing approaches to RefVSR essentially attempt to align the reference and the input sequence, in the presence of resolution gap and long temporal range. However, they either ignore temporal structure within the input sequence, or suffer accumulative alignment errors. To address these issues, we propose EFENet to exploit simultaneously the visual cues contained in the HR reference and the temporal information contained in the LR sequence. EFENet first globally estimates cross-scale flow between the reference and each LR frame. Then our novel flow refinement module of EFENet refines the flow regarding the furthest frame using all the estimated flows, which leverages the global temporal information within the sequence and therefore effectively reduces the alignment errors. We provide comprehensive evaluations to validate the strengths of our approach, and to demonstrate that the proposed framework outperforms the state-of-the-art methods. Code is available at https://github.com/IndigoPurple/EFENet., Comment: 12 pages, 6 figures

In this work, we address the problem of unsupervised domain adaptation for person re-ID where annotations are available for the source domain but not for target. Previous methods typically follow a two-stage optimization pipeline, where the network is first pre-trained on source and then fine-tuned on target with pseudo labels created by feature clustering. Such methods sustain two main limitations. (1) The label noise may hinder the learning of discriminative features for recognizing target classes. (2) The domain gap may hinder knowledge transferring from source to target. We propose three types of technical schemes to alleviate these issues. First, we propose a cluster-wise contrastive learning algorithm (CCL) by iterative optimization of feature learning and cluster refinery to learn noise-tolerant representations in the unsupervised manner. Second, we adopt a progressive domain adaptation (PDA) strategy to gradually mitigate the domain gap between source and target data. Third, we propose Fourier augmentation (FA) for further maximizing the class separability of re-ID models by imposing extra constraints in the Fourier space. We observe that these proposed schemes are capable of facilitating the learning of discriminative feature representations. Experiments demonstrate that our method consistently achieves notable improvements over the state-of-the-art unsupervised re-ID methods on multiple benchmarks, e.g., surpassing MMT largely by 8.1\%, 9.9\%, 11.4\% and 11.1\% mAP on the Market-to-Duke, Duke-to-Market, Market-to-MSMT and Duke-to-MSMT tasks, respectively., Comment: ICCV 2021

Tracking objects of interest in a video is one of the most popular and widely applicable problems in computer vision. However, with the years, a Cambrian explosion of use cases and benchmarks has fragmented the problem in a multitude of different experimental setups. As a consequence, the literature has fragmented too, and now novel approaches proposed by the community are usually specialised to fit only one specific setup. To understand to what extent this specialisation is necessary, in this work we present UniTrack, a solution to address five different tasks within the same framework. UniTrack consists of a single and task-agnostic appearance model, which can be learned in a supervised or self-supervised fashion, and multiple ``heads'' that address individual tasks and do not require training. We show how most tracking tasks can be solved within this framework, and that the same appearance model can be successfully used to obtain results that are competitive against specialised methods for most of the tasks considered. The framework also allows us to analyse appearance models obtained with the most recent self-supervised methods, thus extending their evaluation and comparison to a larger variety of important problems., Comment: To appear at NeurIPS 2021

Detection in large-scale scenes is a challenging problem due to small objects and extreme scale variation. It is essential to focus on the image regions of small objects. In this paper, we propose a novel Adaptive Zoom (AdaZoom) network as a selective magnifier with flexible shape and focal length to adaptively zoom the focus regions for object detection. Based on policy gradient, we construct a reinforcement learning framework for focus region generation, with the reward formulated by object distributions. The scales and aspect ratios of the generated regions are adaptive to the scales and distribution of objects inside. We apply variable magnification according to the scale of the region for adaptive multi-scale detection. We further propose collaborative training to complementarily promote the performance of AdaZoom and the detection network. To validate the effectiveness, we conduct extensive experiments on VisDrone2019, UAVDT, and DOTA datasets. The experiments show AdaZoom brings a consistent and significant improvement over different detection networks, achieving state-of-the-art performance on these datasets, especially outperforming the existing methods by AP of 4.64% on Vis-Drone2019.

Recently unsupervised domain adaptation for the semantic segmentation task has become more and more popular due to high-cost of pixel-level annotation on real-world images. However, most domain adaptation methods are only restricted to single-source-single-target pair, and can not be directly extended to multiple target domains. In this work, we propose a collaborative learning framework to achieve unsupervised multi-target domain adaptation. An unsupervised domain adaptation expert model is first trained for each source-target pair and is further encouraged to collaborate with each other through a bridge built between different target domains. These expert models are further improved by adding the regularization of making the consistent pixel-wise prediction for each sample with the same structured context. To obtain a single model that works across multiple target domains, we propose to simultaneously learn a student model which is trained to not only imitate the output of each expert on the corresponding target domain, but also to pull different expert close to each other with regularization on their weights. Extensive experiments demonstrate that the proposed method can effectively exploit rich structured information contained in both labeled source domain and multiple unlabeled target domains. Not only does it perform well across multiple target domains but also performs favorably against state-of-the-art unsupervised domain adaptation methods specially trained on a single source-target pair, Comment: CVPR 2021

Learning pyramidal feature representations is crucial for recognizing object instances at different scales. Feature Pyramid Network (FPN) is the classic architecture to build a feature pyramid with high-level semantics throughout. However, intrinsic defects in feature extraction and fusion inhibit FPN from further aggregating more discriminative features. In this work, we propose Attention Aggregation based Feature Pyramid Network (A^2-FPN), to improve multi-scale feature learning through attention-guided feature aggregation. In feature extraction, it extracts discriminative features by collecting-distributing multi-level global context features, and mitigates the semantic information loss due to drastically reduced channels. In feature fusion, it aggregates complementary information from adjacent features to generate location-wise reassembly kernels for content-aware sampling, and employs channel-wise reweighting to enhance the semantic consistency before element-wise addition. A^2-FPN shows consistent gains on different instance segmentation frameworks. By replacing FPN with A^2-FPN in Mask R-CNN, our model boosts the performance by 2.1% and 1.6% mask AP when using ResNet-50 and ResNet-101 as backbone, respectively. Moreover, A^2-FPN achieves an improvement of 2.0% and 1.4% mask AP when integrated into the strong baselines such as Cascade Mask R-CNN and Hybrid Task Cascade., Comment: CVPR2021

In this paper, we delve into semi-supervised object detection where unlabeled images are leveraged to break through the upper bound of fully-supervised object detection models. Previous semi-supervised methods based on pseudo labels are severely degenerated by noise and prone to overfit to noisy labels, thus are deficient in learning different unlabeled knowledge well. To address this issue, we propose a data-uncertainty guided multi-phase learning method for semi-supervised object detection. We comprehensively consider divergent types of unlabeled images according to their difficulty levels, utilize them in different phases and ensemble models from different phases together to generate ultimate results. Image uncertainty guided easy data selection and region uncertainty guided RoI Re-weighting are involved in multi-phase learning and enable the detector to concentrate on more certain knowledge. Through extensive experiments on PASCAL VOC and MS COCO, we demonstrate that our method behaves extraordinarily compared to baseline approaches and outperforms them by a large margin, more than 3% on VOC and 2% on COCO., Comment: Accepted by CVPR 2021

Video super-resolution plays an important role in surveillance video analysis and ultra-high-definition video display, which has drawn much attention in both the research and industrial communities. Although many deep learning-based VSR methods have been proposed, it is hard to directly compare these methods since the different loss functions and training datasets have a significant impact on the super-resolution results. In this work, we carefully study and compare three temporal modeling methods (2D CNN with early fusion, 3D CNN with slow fusion and Recurrent Neural Network) for video super-resolution. We also propose a novel Recurrent Residual Network (RRN) for efficient video super-resolution, where residual learning is utilized to stabilize the training of RNN and meanwhile to boost the super-resolution performance. Extensive experiments show that the proposed RRN is highly computational efficiency and produces temporal consistent VSR results with finer details than other temporal modeling methods. Besides, the proposed method achieves state-of-the-art results on several widely used benchmarks., Comment: BMVC 2020

Most video super-resolution methods super-resolve a single reference frame with the help of neighboring frames in a temporal sliding window. They are less efficient compared to the recurrent-based methods. In this work, we propose a novel recurrent video super-resolution method which is both effective and efficient in exploiting previous frames to super-resolve the current frame. It divides the input into structure and detail components which are fed to a recurrent unit composed of several proposed two-stream structure-detail blocks. In addition, a hidden state adaptation module that allows the current frame to selectively use information from hidden state is introduced to enhance its robustness to appearance change and error accumulation. Extensive ablation study validate the effectiveness of the proposed modules. Experiments on several benchmark datasets demonstrate the superior performance of the proposed method compared to state-of-the-art methods on video super-resolution., Comment: ECCV 2020

Video super-resolution, which aims at producing a high-resolution video from its corresponding low-resolution version, has recently drawn increasing attention. In this work, we propose a novel method that can effectively incorporate temporal information in a hierarchical way. The input sequence is divided into several groups, with each one corresponding to a kind of frame rate. These groups provide complementary information to recover missing details in the reference frame, which is further integrated with an attention module and a deep intra-group fusion module. In addition, a fast spatial alignment is proposed to handle videos with large motion. Extensive results demonstrate the capability of the proposed model in handling videos with various motion. It achieves favorable performance against state-of-the-art methods on several benchmark datasets., Comment: CVPR 2020

This paper proposes a self-supervised learning method for the person re-identification (re-ID) problem, where existing unsupervised methods usually rely on pseudo labels, such as those from video tracklets or clustering. A potential drawback of using pseudo labels is that errors may accumulate and it is challenging to estimate the number of pseudo IDs. We introduce a different unsupervised method that allows us to learn pedestrian embeddings from raw videos, without resorting to pseudo labels. The goal is to construct a self-supervised pretext task that matches the person re-ID objective. Inspired by the \emph{data association} concept in multi-object tracking, we propose the \textbf{Cyc}le \textbf{As}sociation (\textbf{CycAs}) task: after performing data association between a pair of video frames forward and then backward, a pedestrian instance is supposed to be associated to itself. To fulfill this goal, the model must learn a meaningful representation that can well describe correspondences between instances in frame pairs. We adapt the discrete association process to a differentiable form, such that end-to-end training becomes feasible. Experiments are conducted in two aspects: We first compare our method with existing unsupervised re-ID methods on seven benchmarks and demonstrate CycAs' superiority. Then, to further validate the practical value of CycAs in real-world applications, we perform training on self-collected videos and report promising performance on standard test sets., Comment: Accepted to ECCV 2020

Multi-target multi-camera tracking (MTMCT) systems track targets across cameras. Due to the continuity of target trajectories, tracking systems usually restrict their data association within a local neighborhood. In single camera tracking, local neighborhood refers to consecutive frames; in multi-camera tracking, it refers to neighboring cameras that the target may appear successively. For similarity estimation, tracking systems often adopt appearance features learned from the re-identification (re-ID) perspective. Different from tracking, re-ID usually does not have access to the trajectory cues that can limit the search space to a local neighborhood. Due to its global matching property, the re-ID perspective requires to learn global appearance features. We argue that the mismatch between the local matching procedure in tracking and the global nature of re-ID appearance features may compromise MTMCT performance. To fit the local matching procedure in MTMCT, in this work, we introduce locality aware appearance metric (LAAM). Specifically, we design an intra-camera metric for single camera tracking, and an inter-camera metric for multi-camera tracking. Both metrics are trained with data pairs sampled from their corresponding local neighborhoods, as opposed to global sampling in the re-ID perspective. We show that the locally learned metrics can be successfully applied on top of several globally learned re-ID features. With the proposed method, we report new state-of-the-art performance on the DukeMTMC dataset, and a substantial improvement on the CityFlow dataset.

Modern multiple object tracking (MOT) systems usually follow the \emph{tracking-by-detection} paradigm. It has 1) a detection model for target localization and 2) an appearance embedding model for data association. Having the two models separately executed might lead to efficiency problems, as the running time is simply a sum of the two steps without investigating potential structures that can be shared between them. Existing research efforts on real-time MOT usually focus on the association step, so they are essentially real-time association methods but not real-time MOT system. In this paper, we propose an MOT system that allows target detection and appearance embedding to be learned in a shared model. Specifically, we incorporate the appearance embedding model into a single-shot detector, such that the model can simultaneously output detections and the corresponding embeddings. We further propose a simple and fast association method that works in conjunction with the joint model. In both components the computation cost is significantly reduced compared with former MOT systems, resulting in a neat and fast baseline for future follow-ups on real-time MOT algorithm design. To our knowledge, this work reports the first (near) real-time MOT system, with a running speed of 22 to 40 FPS depending on the input resolution. Meanwhile, its tracking accuracy is comparable to the state-of-the-art trackers embodying separate detection and embedding (SDE) learning ($64.4\%$ MOTA \vs $66.1\%$ MOTA on MOT-16 challenge). Code and models are available at \url{https://github.com/Zhongdao/Towards-Realtime-MOT}., Comment: Accepted to ECCV 2020

This paper addresses the problem of Monocular Depth Estimation (MDE). Existing approaches on MDE usually model it as a pixel-level regression problem, ignoring the underlying geometry property. We empirically find this may result in sub-optimal solution: while the predicted depth map presents small loss value in one specific view, it may exhibit large loss if viewed in different directions. In this paper, inspired by multi-view stereo (MVS), we propose an Adversarial View-Consistent Learning (AVCL) framework to force the estimated depth map to be all reasonable viewed from multiple views. To this end, we first design a differentiable depth map warping operation, which is end-to-end trainable, and then propose a pose generator to generate novel views for a given image in an adversarial manner. Collaborating with the differentiable depth map warping operation, the pose generator encourages the depth estimation network to learn from hard views, hence produce view-consistent depth maps . We evaluate our method on NYU Depth V2 dataset and the experimental results show promising performance gain upon state-of-the-art MDE approaches., Comment: BMVC 2019 Spotlight

The softmax loss and its variants are widely used as objectives for embedding learning, especially in applications like face recognition. However, the intra- and inter-class objectives in the softmax loss are entangled, therefore a well-optimized inter-class objective leads to relaxation on the intra-class objective, and vice versa. In this paper, we propose to dissect the softmax loss into independent intra- and inter-class objective (D-Softmax). With D-Softmax as objective, we can have a clear understanding of both the intra- and inter-class objective, therefore it is straightforward to tune each part to the best state. Furthermore, we find the computation of the inter-class objective is redundant and propose two sampling-based variants of D-Softmax to reduce the computation cost. Training with regular-scale data, experiments in face verification show D-Softmax is favorably comparable to existing losses such as SphereFace and ArcFace. Training with massive-scale data, experiments show the fast variants of D-Softmax significantly accelerates the training process (such as 64x) with only a minor sacrifice in performance, outperforming existing acceleration methods of softmax in terms of both performance and efficiency., Comment: Accepted to AAAI-2020, Oral presentation

Generic object detection is one of the most fundamental problems in computer vision, yet it is difficult to provide all the bounding-box-level annotations aiming at large-scale object detection for thousands of categories. In this paper, we present a novel cross-supervised learning pipeline for large-scale object detection, denoted as CS-R-FCN. First, we propose to utilize the data flow of image-level annotated images in the fully-supervised two-stage object detection framework, leading to cross-supervised learning combining bounding-box-level annotated data and image-level annotated data. Second, we introduce a semantic aggregation strategy utilizing the relationships among the cross-supervised categories to reduce the unreasonable mutual inhibition effects during the feature learning. Experimental results show that the proposed CS-R-FCN improves the mAP by a large margin compared to previous related works.

Video-based person re-identification has drawn massive attention in recent years due to its extensive applications in video surveillance. While deep learning-based methods have led to significant progress, these methods are limited by ineffectively using complementary information, which is blamed on necessary data augmentation in the training process. Data augmentation has been widely used to mitigate the over-fitting trap and improve the ability of network representation. However, the previous methods adopt image-based data augmentation scheme to individually process the input frames, which corrupts the complementary information between consecutive frames and causes performance degradation. Extensive experiments on three benchmark datasets demonstrate that our framework outperforms the most recent state-of-the-art methods. We also perform cross-dataset validation to prove the generality of our method., Comment: ICIP 2020

Object detection has been a challenging task in computer vision. Although significant progress has been made in object detection with deep neural networks, the attention mechanism is far from development. In this paper, we propose the hybrid attention mechanism for single-stage object detection. First, we present the modules of spatial attention, channel attention and aligned attention for single-stage object detection. In particular, stacked dilated convolution layers with symmetrically fixed rates are constructed to learn spatial attention. The channel attention is proposed with the cross-level group normalization and squeeze-and-excitation module. Aligned attention is constructed with organized deformable filters. Second, the three kinds of attention are unified to construct the hybrid attention mechanism. We then embed the hybrid attention into Retina-Net and propose the efficient single-stage HAR-Net for object detection. The attention modules and the proposed HAR-Net are evaluated on the COCO detection dataset. Experiments demonstrate that hybrid attention can significantly improve the detection accuracy and the HAR-Net can achieve the state-of-the-art 45.8\% mAP, outperform existing single-stage object detectors.