Search

While deep neural networks have achieved remarkable performance, they tend to lack transparency in prediction. The pursuit of greater interpretability in neural networks often results in a degradation of their original performance. Some works strive to improve both interpretability and performance, but they primarily depend on meticulously imposed conditions. In this paper, we propose a simple yet effective framework that acquires more explainable activation heatmaps and simultaneously increase the model performance, without the need for any extra supervision. Specifically, our concise framework introduces a new metric, i.e., explanation consistency, to reweight the training samples adaptively in model learning. The explanation consistency metric is utilized to measure the similarity between the model's visual explanations of the original samples and those of semantic-preserved adversarial samples, whose background regions are perturbed by using image adversarial attack techniques. Our framework then promotes the model learning by paying closer attention to those training samples with a high difference in explanations (i.e., low explanation consistency), for which the current model cannot provide robust interpretations. Comprehensive experimental results on various benchmarks demonstrate the superiority of our framework in multiple aspects, including higher recognition accuracy, greater data debiasing capability, stronger network robustness, and more precise localization ability on both regular networks and interpretable networks. We also provide extensive ablation studies and qualitative analyses to unveil the detailed contribution of each component., Comment: To appear in IEEE Transactions on Multimedia

Amidst the rapid advancements in generative language models, the investigation of how training data shapes the performance of GPT models is still emerging. This paper presents GPTfluence, a novel approach that leverages a featurized simulation to assess the impact of training examples on the training dynamics of GPT models. Our approach not only traces the influence of individual training instances on performance trajectories, such as loss and other key metrics, on targeted test points but also enables a comprehensive comparison with existing methods across various training scenarios in GPT models, ranging from 14 million to 2.8 billion parameters, across a range of downstream tasks. Contrary to earlier methods that struggle with generalization to new data, GPTfluence introduces a parameterized simulation of training dynamics, demonstrating robust generalization capabilities to unseen training data. This adaptability is evident across both fine-tuning and instruction-tuning scenarios, spanning tasks in natural language understanding and generation. We will make our code and data publicly available.

Neural Radiance Fields (NeRF) have garnered remarkable success in novel view synthesis. Nonetheless, the task of generating high-quality images for novel views persists as a critical challenge. While the existing efforts have exhibited commendable progress, capturing intricate details, enhancing textures, and achieving superior Peak Signal-to-Noise Ratio (PSNR) metrics warrant further focused attention and advancement. In this work, we propose NeRF-VPT, an innovative method for novel view synthesis to address these challenges. Our proposed NeRF-VPT employs a cascading view prompt tuning paradigm, wherein RGB information gained from preceding rendering outcomes serves as instructive visual prompts for subsequent rendering stages, with the aspiration that the prior knowledge embedded in the prompts can facilitate the gradual enhancement of rendered image quality. NeRF-VPT only requires sampling RGB data from previous stage renderings as priors at each training stage, without relying on extra guidance or complex techniques. Thus, our NeRF-VPT is plug-and-play and can be readily integrated into existing methods. By conducting comparative analyses of our NeRF-VPT against several NeRF-based approaches on demanding real-scene benchmarks, such as Realistic Synthetic 360, Real Forward-Facing, Replica dataset, and a user-captured dataset, we substantiate that our NeRF-VPT significantly elevates baseline performance and proficiently generates more high-quality novel view images than all the compared state-of-the-art methods. Furthermore, the cascading learning of NeRF-VPT introduces adaptability to scenarios with sparse inputs, resulting in a significant enhancement of accuracy for sparse-view novel view synthesis. The source code and dataset are available at \url{https://github.com/Freedomcls/NeRF-VPT}., Comment: AAAI 2024

Recently, Vision Transformer has achieved great success in recovering missing details in low-resolution sequences, i.e., the video super-resolution (VSR) task. Despite its superiority in VSR accuracy, the heavy computational burden as well as the large memory footprint hinder the deployment of Transformer-based VSR models on constrained devices. In this paper, we address the above issue by proposing a novel feature-level masked processing framework: VSR with Masked Intra and inter frame Attention (MIA-VSR). The core of MIA-VSR is leveraging feature-level temporal continuity between adjacent frames to reduce redundant computations and make more rational use of previously enhanced SR features. Concretely, we propose an intra-frame and inter-frame attention block which takes the respective roles of past features and input features into consideration and only exploits previously enhanced features to provide supplementary information. In addition, an adaptive block-wise mask prediction module is developed to skip unimportant computations according to feature similarity between adjacent frames. We conduct detailed ablation studies to validate our contributions and compare the proposed method with recent state-of-the-art VSR approaches. The experimental results demonstrate that MIA-VSR improves the memory and computation efficiency over state-of-the-art methods, without trading off PSNR accuracy. The code is available at https://github.com/LabShuHangGU/MIA-VSR., Comment: Accepted by CVPR 2024

Speech-driven 3D facial animation aims to synthesize vivid facial animations that accurately synchronize with speech and match the unique speaking style. However, existing works primarily focus on achieving precise lip synchronization while neglecting to model the subject-specific speaking style, often resulting in unrealistic facial animations. To the best of our knowledge, this work makes the first attempt to explore the coupled information between the speaking style and the semantic content in facial motions. Specifically, we introduce an innovative speaking style disentanglement method, which enables arbitrary-subject speaking style encoding and leads to a more realistic synthesis of speech-driven facial animations. Subsequently, we propose a novel framework called \textbf{Mimic} to learn disentangled representations of the speaking style and content from facial motions by building two latent spaces for style and content, respectively. Moreover, to facilitate disentangled representation learning, we introduce four well-designed constraints: an auxiliary style classifier, an auxiliary inverse classifier, a content contrastive loss, and a pair of latent cycle losses, which can effectively contribute to the construction of the identity-related style space and semantic-related content space. Extensive qualitative and quantitative experiments conducted on three publicly available datasets demonstrate that our approach outperforms state-of-the-art methods and is capable of capturing diverse speaking styles for speech-driven 3D facial animation. The source code and supplementary video are publicly available at: https://zeqing-wang.github.io/Mimic/, Comment: 7 pages, 6 figures, accepted by AAAI-24

Recently, several methods have been proposed to augment large Vision Language Models (VLMs) for Visual Question Answering (VQA) simplicity by incorporating external knowledge from knowledge bases or visual clues derived from question decomposition. Although having achieved promising results, these methods still suffer from the challenge that VLMs cannot inherently understand the incorporated knowledge and might fail to generate the optimal answers. Contrarily, human cognition engages visual questions through a top-down reasoning process, systematically exploring relevant issues to derive a comprehensive answer. This not only facilitates an accurate answer but also provides a transparent rationale for the decision-making pathway. Motivated by this cognitive mechanism, we introduce a novel, explainable multi-agent collaboration framework designed to imitate human-like top-down reasoning by leveraging the expansive knowledge of Large Language Models (LLMs). Our framework comprises three agents, i.e., Responder, Seeker, and Integrator, each contributing uniquely to the top-down reasoning process. The VLM-based Responder generates the answer candidates for the question and gives responses to other issues. The Seeker, primarily based on LLM, identifies relevant issues related to the question to inform the Responder and constructs a Multi-View Knowledge Base (MVKB) for the given visual scene by leveraging the understanding capabilities of LLM. The Integrator agent combines information from the Seeker and the Responder to produce the final VQA answer. Through this collaboration mechanism, our framework explicitly constructs an MVKB for a specific visual scene and reasons answers in a top-down reasoning process. Extensive and comprehensive evaluations on diverse VQA datasets and VLMs demonstrate the superior applicability and interpretability of our framework over the existing compared methods., Comment: 16 pages, 5 figures

The class-agnostic counting (CAC) task has recently been proposed to solve the problem of counting all objects of an arbitrary class with several exemplars given in the input image. To address this challenging task, existing leading methods all resort to density map regression, which renders them impractical for downstream tasks that require object locations and restricts their ability to well explore the scale information of exemplars for supervision. To address the limitations, we propose a novel localization-based CAC approach, termed Scale-modulated Query and Localization Network (SQLNet). It fully explores the scales of exemplars in both the query and localization stages and achieves effective counting by accurately locating each object and predicting its approximate size. Specifically, during the query stage, rich discriminative representations of the target class are acquired by the Hierarchical Exemplars Collaborative Enhancement (HECE) module from the few exemplars through multi-scale exemplar cooperation with equifrequent size prompt embedding. These representations are then fed into the Exemplars-Unified Query Correlation (EUQC) module to interact with the query features in a unified manner and produce the correlated query tensor. In the localization stage, the Scale-aware Multi-head Localization (SAML) module utilizes the query tensor to predict the confidence, location, and size of each potential object. Moreover, a scale-aware localization loss is introduced, which exploits flexible location associations and exemplar scales for supervision to optimize the model performance. Extensive experiments demonstrate that SQLNet outperforms state-of-the-art methods on popular CAC benchmarks, achieving excellent performance not only in counting accuracy but also in localization and bounding box generation. Our codes will be available at https://github.com/HCPLab-SYSU/SQLNet, Comment: 13 pages

As an interpretable and universal neuro-symbolic paradigm based on Large Language Models, visual programming (VisualProg) can execute compositional visual tasks without training, but its performance is markedly inferior compared to task-specific supervised learning models. To increase its practicality, the performance of VisualProg on specific tasks needs to be improved. However, the non-differentiability of VisualProg limits the possibility of employing the fine-tuning strategy on specific tasks to achieve further improvements. In our analysis, we discovered that significant performance issues in VisualProg's execution originated from errors made by the sub-modules at corresponding visual sub-task steps. To address this, we propose ``VisualProg Distiller", a method of supplementing and distilling process knowledge to optimize the performance of each VisualProg sub-module on decoupled visual sub-tasks, thus enhancing the overall task performance. Specifically, we choose an end-to-end model that is well-performed on the given task as the teacher and further distill the knowledge of the teacher into the invoked visual sub-modules step-by-step based on the execution flow of the VisualProg-generated programs. In this way, our method is capable of facilitating the fine-tuning of the non-differentiable VisualProg frameworks effectively. Extensive and comprehensive experimental evaluations demonstrate that our method can achieve a substantial performance improvement of VisualProg, and outperforms all the compared state-of-the-art methods by large margins. Furthermore, to provide valuable process supervision for the GQA task, we construct a large-scale dataset by utilizing the distillation process of our method., Comment: 12 pages, 10 figures, 6 tables

Despite advancements in LLMs, knowledge-based reasoning remains a longstanding issue due to the fragility of knowledge recall and inference. Existing methods primarily encourage LLMs to autonomously plan and solve problems or to extensively sample reasoning chains without addressing the conceptual and inferential fallacies. Attempting to alleviate inferential fallacies and drawing inspiration from multi-agent collaboration, we present a framework to increase faithfulness and causality for knowledge-based reasoning. Specifically, we propose to employ multiple intelligent agents (i.e., reasoners and an evaluator) to work collaboratively in a reasoning-and-consensus paradigm for elevated reasoning faithfulness. The reasoners focus on providing solutions with human-like causality to solve open-domain problems. On the other hand, the \textit{evaluator} agent scrutinizes if a solution is deducible from a non-causal perspective and if it still holds when challenged by a counterfactual candidate. According to the extensive and comprehensive evaluations on a variety of knowledge reasoning tasks (e.g., science question answering and commonsense reasoning), our framework outperforms all compared state-of-the-art approaches by large margins., Comment: 8 pages, 3 figures. 4 tables

Video self-supervised learning is a challenging task, which requires significant expressive power from the model to leverage rich spatial-temporal knowledge and generate effective supervisory signals from large amounts of unlabeled videos. However, existing methods fail to increase the temporal diversity of unlabeled videos and ignore elaborately modeling multi-scale temporal dependencies in an explicit way. To overcome these limitations, we take advantage of the multi-scale temporal dependencies within videos and proposes a novel video self-supervised learning framework named Temporal Contrastive Graph Learning (TCGL), which jointly models the inter-snippet and intra-snippet temporal dependencies for temporal representation learning with a hybrid graph contrastive learning strategy. Specifically, a Spatial-Temporal Knowledge Discovering (STKD) module is first introduced to extract motion-enhanced spatial-temporal representations from videos based on the frequency domain analysis of discrete cosine transform. To explicitly model multi-scale temporal dependencies of unlabeled videos, our TCGL integrates the prior knowledge about the frame and snippet orders into graph structures, i.e., the intra-/inter- snippet Temporal Contrastive Graphs (TCG). Then, specific contrastive learning modules are designed to maximize the agreement between nodes in different graph views. To generate supervisory signals for unlabeled videos, we introduce an Adaptive Snippet Order Prediction (ASOP) module which leverages the relational knowledge among video snippets to learn the global context representation and recalibrate the channel-wise features adaptively. Experimental results demonstrate the superiority of our TCGL over the state-of-the-art methods on large-scale action recognition and video retrieval benchmarks.The code is publicly available at https://github.com/YangLiu9208/TCGL., Comment: This work has been published in IEEE Transactions on Image Processing. The code is publicly available at https://github.com/YangLiu9208/TCGL. arXiv admin note: substantial text overlap with arXiv:2101.00820

Aiming at recognizing the samples from novel categories with few reference samples, few-shot learning (FSL) is a challenging problem. We found that the existing works often build their few-shot model based on the image-level feature by mixing all local-level features, which leads to the discriminative location bias and information loss in local details. To tackle the problem, this paper returns the perspective to the local-level feature and proposes a series of local-level strategies. Specifically, we present (a) a local-agnostic training strategy to avoid the discriminative location bias between the base and novel categories, (b) a novel local-level similarity measure to capture the accurate comparison between local-level features, and (c) a local-level knowledge transfer that can synthesize different knowledge transfers from the base category according to different location features. Extensive experiments justify that our proposed local-level strategies can significantly boost the performance and achieve 2.8%-7.2% improvements over the baseline across different benchmark datasets, which also achieves state-of-the-art accuracy., Comment: 5 pages, 4 figures, submitted to ICASSP 2022

We propose CX-ToM, short for counterfactual explanations with theory-of mind, a new explainable AI (XAI) framework for explaining decisions made by a deep convolutional neural network (CNN). In contrast to the current methods in XAI that generate explanations as a single shot response, we pose explanation as an iterative communication process, i.e. dialog, between the machine and human user. More concretely, our CX-ToM framework generates sequence of explanations in a dialog by mediating the differences between the minds of machine and human user. To do this, we use Theory of Mind (ToM) which helps us in explicitly modeling human's intention, machine's mind as inferred by the human as well as human's mind as inferred by the machine. Moreover, most state-of-the-art XAI frameworks provide attention (or heat map) based explanations. In our work, we show that these attention based explanations are not sufficient for increasing human trust in the underlying CNN model. In CX-ToM, we instead use counterfactual explanations called fault-lines which we define as follows: given an input image I for which a CNN classification model M predicts class c_pred, a fault-line identifies the minimal semantic-level features (e.g., stripes on zebra, pointed ears of dog), referred to as explainable concepts, that need to be added to or deleted from I in order to alter the classification category of I by M to another specified class c_alt. We argue that, due to the iterative, conceptual and counterfactual nature of CX-ToM explanations, our framework is practical and more natural for both expert and non-expert users to understand the internal workings of complex deep learning models. Extensive quantitative and qualitative experiments verify our hypotheses, demonstrating that our CX-ToM significantly outperforms the state-of-the-art explainable AI models., Comment: Accepted by iScience Cell Press Journal 2021. arXiv admin note: text overlap with arXiv:1909.06907

Self-supervised learning (especially contrastive learning) has attracted great interest due to its huge potential in learning discriminative representations in an unsupervised manner. Despite the acknowledged successes, existing contrastive learning methods suffer from very low learning efficiency, e.g., taking about ten times more training epochs than supervised learning for comparable recognition accuracy. In this paper, we reveal two contradictory phenomena in contrastive learning that we call under-clustering and over-clustering problems, which are major obstacles to learning efficiency. Under-clustering means that the model cannot efficiently learn to discover the dissimilarity between inter-class samples when the negative sample pairs for contrastive learning are insufficient to differentiate all the actual object classes. Over-clustering implies that the model cannot efficiently learn features from excessive negative sample pairs, forcing the model to over-cluster samples of the same actual classes into different clusters. To simultaneously overcome these two problems, we propose a novel self-supervised learning framework using a truncated triplet loss. Precisely, we employ a triplet loss tending to maximize the relative distance between the positive pair and negative pairs to address the under-clustering problem; and we construct the negative pair by selecting a negative sample deputy from all negative samples to avoid the over-clustering problem, guaranteed by the Bernoulli Distribution model. We extensively evaluate our framework in several large-scale benchmarks (e.g., ImageNet, SYSU-30k, and COCO). The results demonstrate our model's superiority (e.g., the learning efficiency) over the latest state-of-the-art methods by a clear margin. Codes available at: https://github.com/wanggrun/triplet ., Comment: ICCV 2021 paper, oral presentation

Attempt to fully discover the temporal diversity and chronological characteristics for self-supervised video representation learning, this work takes advantage of the temporal dependencies within videos and further proposes a novel self-supervised method named Temporal Contrastive Graph Learning (TCGL). In contrast to the existing methods that ignore modeling elaborate temporal dependencies, our TCGL roots in a hybrid graph contrastive learning strategy to jointly regard the inter-snippet and intra-snippet temporal dependencies as self-supervision signals for temporal representation learning. To model multi-scale temporal dependencies, our TCGL integrates the prior knowledge about the frame and snippet orders into graph structures, i.e., the intra-/inter- snippet temporal contrastive graphs. By randomly removing edges and masking nodes of the intra-snippet graphs or inter-snippet graphs, our TCGL can generate different correlated graph views. Then, specific contrastive learning modules are designed to maximize the agreement between nodes in different views. To adaptively learn the global context representation and recalibrate the channel-wise features, we introduce an adaptive video snippet order prediction module, which leverages the relational knowledge among video snippets to predict the actual snippet orders. Experimental results demonstrate the superiority of our TCGL over the state-of-the-art methods on large-scale action recognition and video retrieval benchmarks., Comment: 12 pages, 6 figures

Though beneficial for encouraging the Visual Question Answering (VQA) models to discover the underlying knowledge by exploiting the input-output correlation beyond image and text contexts, the existing knowledge VQA datasets are mostly annotated in a crowdsource way, e.g., collecting questions and external reasons from different users via the internet. In addition to the challenge of knowledge reasoning, how to deal with the annotator bias also remains unsolved, which often leads to superficial over-fitted correlations between questions and answers. To address this issue, we propose a novel dataset named Knowledge-Routed Visual Question Reasoning for VQA model evaluation. Considering that a desirable VQA model should correctly perceive the image context, understand the question, and incorporate its learned knowledge, our proposed dataset aims to cutoff the shortcut learning exploited by the current deep embedding models and push the research boundary of the knowledge-based visual question reasoning. Specifically, we generate the question-answer pair based on both the Visual Genome scene graph and an external knowledge base with controlled programs to disentangle the knowledge from other biases. The programs can select one or two triplets from the scene graph or knowledge base to push multi-step reasoning, avoid answer ambiguity, and balanced the answer distribution. In contrast to the existing VQA datasets, we further imply the following two major constraints on the programs to incorporate knowledge reasoning: i) multiple knowledge triplets can be related to the question, but only one knowledge relates to the image object. This can enforce the VQA model to correctly perceive the image instead of guessing the knowledge based on the given question solely; ii) all questions are based on different knowledge, but the candidate answers are the same for both the training and test sets., Comment: To appear in TNNLS 2021. Considering that a desirable VQA model should correctly perceive the image context, understand the question, and incorporate its learned knowledge, our proposed dataset aims to cutoff the shortcut learning exploited by the current deep embedding models and push the research boundary of the knowledge-based visual question reasoning

Although deep reinforcement learning (RL) has been successfully applied to a variety of robotic control tasks, it's still challenging to apply it to real-world tasks, due to the poor sample efficiency. Attempting to overcome this shortcoming, several works focus on reusing the collected trajectory data during the training by decomposing them into a set of policy-irrelevant discrete transitions. However, their improvements are somewhat marginal since i) the amount of the transitions is usually small, and ii) the value assignment only happens in the joint states. To address these issues, this paper introduces a concise yet powerful method to construct Continuous Transition, which exploits the trajectory information by exploiting the potential transitions along the trajectory. Specifically, we propose to synthesize new transitions for training by linearly interpolating the consecutive transitions. To keep the constructed transitions authentic, we also develop a discriminator to guide the construction process automatically. Extensive experiments demonstrate that our proposed method achieves a significant improvement in sample efficiency on various complex continuous robotic control problems in MuJoCo and outperforms the advanced model-based / model-free RL methods. The source code is available.

Existing vision-based action recognition is susceptible to occlusion and appearance variations, while wearable sensors can alleviate these challenges by capturing human motion with one-dimensional time-series signal. For the same action, the knowledge learned from vision sensors and wearable sensors, may be related and complementary. However, there exists significantly large modality difference between action data captured by wearable-sensor and vision-sensor in data dimension, data distribution and inherent information content. In this paper, we propose a novel framework, named Semantics-aware Adaptive Knowledge Distillation Networks (SAKDN), to enhance action recognition in vision-sensor modality (videos) by adaptively transferring and distilling the knowledge from multiple wearable sensors. The SAKDN uses multiple wearable-sensors as teacher modalities and uses RGB videos as student modality. To preserve local temporal relationship and facilitate employing visual deep learning model, we transform one-dimensional time-series signals of wearable sensors to two-dimensional images by designing a gramian angular field based virtual image generation model. Then, we build a novel Similarity-Preserving Adaptive Multi-modal Fusion Module to adaptively fuse intermediate representation knowledge from different teacher networks. Finally, to fully exploit and transfer the knowledge of multiple well-trained teacher networks to the student network, we propose a novel Graph-guided Semantically Discriminative Mapping loss, which utilizes graph-guided ablation analysis to produce a good visual explanation highlighting the important regions across modalities and concurrently preserving the interrelations of original data. Experimental results on Berkeley-MHAD, UTD-MHAD and MMAct datasets well demonstrate the effectiveness of our proposed SAKDN., Comment: This paper focuses on the sensor-to-vision heterogenous action recognition problem. Code is available at https://github.com/YangLiu9208/SAKDN

Recently, studies of visual question answering have explored various architectures of end-to-end networks and achieved promising results on both natural and synthetic datasets, which require explicitly compositional reasoning. However, it has been argued that these black-box approaches lack interpretability of results, and thus cannot perform well on generalization tasks due to overfitting the dataset bias. In this work, we aim to combine the benefits of both sides and overcome their limitations to achieve an end-to-end interpretable structural reasoning for general images without the requirement of layout annotations. Inspired by the property of a capsule network that can carve a tree structure inside a regular convolutional neural network (CNN), we propose a hierarchical compositional reasoning model called the "Linguistically driven Graph Capsule Network", where the compositional process is guided by the linguistic parse tree. Specifically, we bind each capsule in the lowest layer to bridge the linguistic embedding of a single word in the original question with visual evidence and then route them to the same capsule if they are siblings in the parse tree. This compositional process is achieved by performing inference on a linguistically driven conditional random field (CRF) and is performed across multiple graph capsule layers, which results in a compositional reasoning process inside a CNN. Experiments on the CLEVR dataset, CLEVR compositional generation test, and FigureQA dataset demonstrate the effectiveness and composition generalization ability of our end-to-end model., Comment: Submitted to TPAMI 2020. We have achieved an end-to-end interpretable structural reasoning for general images without the requirement of layout annotations

Medical diagnosis assistant (MDA) aims to build an interactive diagnostic agent to sequentially inquire about symptoms for discriminating diseases. However, since the dialogue records used to build a patient simulator are collected passively, the data might be deteriorated by some task-unrelated biases, such as the preference of the collectors. These biases might hinder the diagnostic agent to capture transportable knowledge from the simulator. This work attempts to address these critical issues in MDA by taking advantage of the causal diagram to identify and resolve two representative non-causal biases, i.e., (i) default-answer bias and (ii) distributional inquiry bias. Specifically, Bias (i) originates from the patient simulator which tries to answer the unrecorded inquiries with some biased default answers. Consequently, the diagnostic agents cannot fully demonstrate their advantages due to the biased answers. To eliminate this bias and inspired by the propensity score matching technique with causal diagram, we propose a propensity-based patient simulator to effectively answer unrecorded inquiry by drawing knowledge from the other records; Bias (ii) inherently comes along with the passively collected data, and is one of the key obstacles for training the agent towards "learning how" rather than "remembering what". For example, within the distribution of training data, if a symptom is highly coupled with a certain disease, the agent might learn to only inquire about that symptom to discriminate that disease, thus might not generalize to the out-of-distribution cases. To this end, we propose a progressive assurance agent, which includes the dual processes accounting for symptom inquiry and disease diagnosis respectively. The inquiry process is driven by the diagnosis process in a top-down manner to inquire about symptoms for enhancing diagnostic confidence., Comment: To appear in TPAMI 2023. In the experiments, our trained agent achieves the new state-of-the-art under various experimental settings and possesses the advantage of sample-efficiency and robustness compared to other existing MDA methods

Multi-Person Tracking (MPT) is often addressed within the detection-to-association paradigm. In such approaches, human detections are first extracted in every frame and person trajectories are then recovered by a procedure of data association (usually offline). However, their performances usually degenerate in presence of detection errors, mutual interactions and occlusions. In this paper, we present a deep learning based MPT approach that learns instance-aware representations of tracked persons and robustly online infers states of the tracked persons. Specifically, we design a multi-branch neural network (MBN), which predicts the classification confidences and locations of all targets by taking a batch of candidate regions as input. In our MBN architecture, each branch (instance-subnet) corresponds to an individual to be tracked and new branches can be dynamically created for handling newly appearing persons. Then based on the output of MBN, we construct a joint association matrix that represents meaningful states of tracked persons (e.g., being tracked or disappearing from the scene) and solve it by using the efficient Hungarian algorithm. Moreover, we allow the instance-subnets to be updated during tracking by online mining hard examples, accounting to person appearance variations over time. We comprehensively evaluate our framework on a popular MPT benchmark, demonstrating its excellent performance in comparison with recent online MPT methods., Comment: accepted by Pattern Recognition

Face hallucination is a domain-specific super-resolution problem that aims to generate a high-resolution (HR) face image from a low-resolution~(LR) input. In contrast to the existing patch-wise super-resolution models that divide a face image into regular patches and independently apply LR to HR mapping to each patch, we implement deep reinforcement learning and develop a novel attention-aware face hallucination (Attention-FH) framework, which recurrently learns to attend a sequence of patches and performs facial part enhancement by fully exploiting the global interdependency of the image. Specifically, our proposed framework incorporates two components: a recurrent policy network for dynamically specifying a new attended region at each time step based on the status of the super-resolved image and the past attended region sequence, and a local enhancement network for selected patch hallucination and global state updating. The Attention-FH model jointly learns the recurrent policy network and local enhancement network through maximizing a long-term reward that reflects the hallucination result with respect to the whole HR image. Extensive experiments demonstrate that our Attention-FH significantly outperforms the state-of-the-art methods on in-the-wild face images with large pose and illumination variations., Comment: To be published in TPAMI

This paper presents a novel adaptively connected neural network (ACNet) to improve the traditional convolutional neural networks (CNNs) {in} two aspects. First, ACNet employs a flexible way to switch global and local inference in processing the internal feature representations by adaptively determining the connection status among the feature nodes (e.g., pixels of the feature maps) \footnote{In a computer vision domain, a node refers to a pixel of a feature map{, while} in {the} graph domain, a node denotes a graph node.}. We can show that existing CNNs, the classical multilayer perceptron (MLP), and the recently proposed non-local network (NLN) \cite{nonlocalnn17} are all special cases of ACNet. Second, ACNet is also capable of handling non-Euclidean data. Extensive experimental analyses on {a variety of benchmarks (i.e.,} ImageNet-1k classification, COCO 2017 detection and segmentation, CUHK03 person re-identification, CIFAR analysis, and Cora document categorization) demonstrate that {ACNet} cannot only achieve state-of-the-art performance but also overcome the limitation of the conventional MLP and CNN \footnote{Corresponding author: Liang Lin (linliang@ieee.org)}. The code is available at \url{https://github.com/wanggrun/Adaptively-Connected-Neural-Networks}., Comment: Accepted by CVPR 2019

Driven by recent computer vision and robotic applications, recovering 3D human poses has become increasingly important and attracted growing interests. In fact, completing this task is quite challenging due to the diverse appearances, viewpoints, occlusions and inherently geometric ambiguities inside monocular images. Most of the existing methods focus on designing some elaborate priors /constraints to directly regress 3D human poses based on the corresponding 2D human pose-aware features or 2D pose predictions. However, due to the insufficient 3D pose data for training and the domain gap between 2D space and 3D space, these methods have limited scalabilities for all practical scenarios (e.g., outdoor scene). Attempt to address this issue, this paper proposes a simple yet effective self-supervised correction mechanism to learn all intrinsic structures of human poses from abundant images. Specifically, the proposed mechanism involves two dual learning tasks, i.e., the 2D-to-3D pose transformation and 3D-to-2D pose projection, to serve as a bridge between 3D and 2D human poses in a type of "free" self-supervision for accurate 3D human pose estimation. The 2D-to-3D pose implies to sequentially regress intermediate 3D poses by transforming the pose representation from the 2D domain to the 3D domain under the sequence-dependent temporal context, while the 3D-to-2D pose projection contributes to refining the intermediate 3D poses by maintaining geometric consistency between the 2D projections of 3D poses and the estimated 2D poses. We further apply our self-supervised correction mechanism to develop a 3D human pose machine, which jointly integrates the 2D spatial relationship, temporal smoothness of predictions and 3D geometric knowledge. Extensive evaluations demonstrate the superior performance and efficiency of our framework over all the compared competing methods., Comment: To appear in IEEE Transactions on Pattern Analysis and Machine Intelligence (T-PAMI), 2019. Our simple yet effective self-supervised correction mechanism to incorporate 3D pose geometric structural information is innovative in the literature, and may also inspire other 3D vision tasks. Please find the code of this project at: http://www.sysu-hcp.net/3d_pose_ssl/

Though quite challenging, leveraging large-scale unlabeled or partially labeled data in learning systems (e.g., model/classifier training) has attracted increasing attentions due to its fundamental importance. To address this problem, many active learning (AL) methods have been proposed that employ up-to-date detectors to retrieve representative minority samples according to predefined confidence or uncertainty thresholds. However, these AL methods cause the detectors to ignore the remaining majority samples (i.e., those with low uncertainty or high prediction confidence). In this work, by developing a principled active sample mining (ASM) framework, we demonstrate that cost-effectively mining samples from these unlabeled majority data is key to training more powerful object detectors while minimizing user effort. Specifically, our ASM framework involves a switchable sample selection mechanism for determining whether an unlabeled sample should be manually annotated via AL or automatically pseudo-labeled via a novel self-learning process. The proposed process can be compatible with mini-batch based training (i.e., using a batch of unlabeled or partially labeled data as a one-time input) for object detection. In addition, a few samples with low-confidence predictions are selected and annotated via AL. Notably, our method is suitable for object categories that are not seen in the unlabeled data during the learning process. Extensive experiments clearly demonstrate that our ASM framework can achieve performance comparable to that of alternative methods but with significantly fewer annotations., Comment: Automatically determining whether an unlabeled sample should be manually annotated or pseudo-labeled via a novel self-learning process (Accepted by TNNLS 2018) The source code is available at http://kezewang.com/codes/ASM_ver1.zip

Though quite challenging, leveraging large-scale unlabeled or partially labeled images in a cost-effective way has increasingly attracted interests for its great importance to computer vision. To tackle this problem, many Active Learning (AL) methods have been developed. However, these methods mainly define their sample selection criteria within a single image context, leading to the suboptimal robustness and impractical solution for large-scale object detection. In this paper, aiming to remedy the drawbacks of existing AL methods, we present a principled Self-supervised Sample Mining (SSM) process accounting for the real challenges in object detection. Specifically, our SSM process concentrates on automatically discovering and pseudo-labeling reliable region proposals for enhancing the object detector via the introduced cross image validation, i.e., pasting these proposals into different labeled images to comprehensively measure their values under different image contexts. By resorting to the SSM process, we propose a new AL framework for gradually incorporating unlabeled or partially labeled data into the model learning while minimizing the annotating effort of users. Extensive experiments on two public benchmarks clearly demonstrate our proposed framework can achieve the comparable performance to the state-of-the-art methods with significantly fewer annotations., Comment: We enabled to mine from unlabeled or partially labeled data to boost object detection (Accepted by CVPR 2018) The source code is available at http://kezewang.com/codes/SSM_CVPR.zip

3D human articulated pose recovery from monocular image sequences is very challenging due to the diverse appearances, viewpoints, occlusions, and also the human 3D pose is inherently ambiguous from the monocular imagery. It is thus critical to exploit rich spatial and temporal long-range dependencies among body joints for accurate 3D pose sequence prediction. Existing approaches usually manually design some elaborate prior terms and human body kinematic constraints for capturing structures, which are often insufficient to exploit all intrinsic structures and not scalable for all scenarios. In contrast, this paper presents a Recurrent 3D Pose Sequence Machine(RPSM) to automatically learn the image-dependent structural constraint and sequence-dependent temporal context by using a multi-stage sequential refinement. At each stage, our RPSM is composed of three modules to predict the 3D pose sequences based on the previously learned 2D pose representations and 3D poses: (i) a 2D pose module extracting the image-dependent pose representations, (ii) a 3D pose recurrent module regressing 3D poses and (iii) a feature adaption module serving as a bridge between module (i) and (ii) to enable the representation transformation from 2D to 3D domain. These three modules are then assembled into a sequential prediction framework to refine the predicted poses with multiple recurrent stages. Extensive evaluations on the Human3.6M dataset and HumanEva-I dataset show that our RPSM outperforms all state-of-the-art approaches for 3D pose estimation., Comment: Published in CVPR 2017

Aiming at improving performance of visual classification in a cost-effective manner, this paper proposes an incremental semi-supervised learning paradigm called Deep Co-Space (DCS). Unlike many conventional semi-supervised learning methods usually performing within a fixed feature space, our DCS gradually propagates information from labeled samples to unlabeled ones along with deep feature learning. We regard deep feature learning as a series of steps pursuing feature transformation, i.e., projecting the samples from a previous space into a new one, which tends to select the reliable unlabeled samples with respect to this setting. Specifically, for each unlabeled image instance, we measure its reliability by calculating the category variations of feature transformation from two different neighborhood variation perspectives, and merged them into an unified sample mining criterion deriving from Hellinger distance. Then, those samples keeping stable correlation to their neighboring samples (i.e., having small category variation in distribution) across the successive feature space transformation, are automatically received labels and incorporated into the model for incrementally training in terms of classification. Our extensive experiments on standard image classification benchmarks (e.g., Caltech-256 and SUN-397) demonstrate that the proposed framework is capable of effectively mining from large-scale unlabeled images, which boosts image classification performance and achieves promising results compared to other semi-supervised learning methods., Comment: To appear in IEEE Transactions on Circuits and Systems for Video Technology (T-CSVT), 2017

Single image super resolution (SR), which refers to reconstruct a higher-resolution (HR) image from the observed low-resolution (LR) image, has received substantial attention due to its tremendous application potentials. Despite the breakthroughs of recently proposed SR methods using convolutional neural networks (CNNs), their generated results usually lack of preserving structural (high-frequency) details. In this paper, regarding global boundary context and residual context as complimentary information for enhancing structural details in image restoration, we develop a contextualized multi-task learning framework to address the SR problem. Specifically, our method first extracts convolutional features from the input LR image and applies one deconvolutional module to interpolate the LR feature maps in a content-adaptive way. Then, the resulting feature maps are fed into two branched sub-networks. During the neural network training, one sub-network outputs salient image boundaries and the HR image, and the other sub-network outputs the local residual map, i.e., the residual difference between the generated HR image and ground-truth image. On several standard benchmarks (i.e., Set5, Set14 and BSD200), our extensive evaluations demonstrate the effectiveness of our SR method on achieving both higher restoration quality and computational efficiency compared with several state-of-the-art SR approaches. The source code and some SR results can be found at: http://hcp.sysu.edu.cn/structure-preserving-image-super-resolution/, Comment: To appear in Transactions on Multimedia 2017

Recent successes in learning-based image classification, however, heavily rely on the large number of annotated training samples, which may require considerable human efforts. In this paper, we propose a novel active learning framework, which is capable of building a competitive classifier with optimal feature representation via a limited amount of labeled training instances in an incremental learning manner. Our approach advances the existing active learning methods in two aspects. First, we incorporate deep convolutional neural networks into active learning. Through the properly designed framework, the feature representation and the classifier can be simultaneously updated with progressively annotated informative samples. Second, we present a cost-effective sample selection strategy to improve the classification performance with less manual annotations. Unlike traditional methods focusing on only the uncertain samples of low prediction confidence, we especially discover the large amount of high confidence samples from the unlabeled set for feature learning. Specifically, these high confidence samples are automatically selected and iteratively assigned pseudo-labels. We thus call our framework "Cost-Effective Active Learning" (CEAL) standing for the two advantages.Extensive experiments demonstrate that the proposed CEAL framework can achieve promising results on two challenging image classification datasets, i.e., face recognition on CACD database [1] and object categorization on Caltech-256 [2]., Comment: Accepted by IEEE Transactions on Circuits and Systems for Video Technology (TCSVT) 2016

This paper aims to develop a novel cost-effective framework for face identification, which progressively maintains a batch of classifiers with the increasing face images of different individuals. By naturally combining two recently rising techniques: active learning (AL) and self-paced learning (SPL), our framework is capable of automatically annotating new instances and incorporating them into training under weak expert re-certification. We first initialize the classifier using a few annotated samples for each individual, and extract image features using the convolutional neural nets. Then, a number of candidates are selected from the unannotated samples for classifier updating, in which we apply the current classifiers ranking the samples by the prediction confidence. In particular, our approach utilizes the high-confidence and low-confidence samples in the self-paced and the active user-query way, respectively. The neural nets are later fine-tuned based on the updated classifiers. Such heuristic implementation is formulated as solving a concise active SPL optimization problem, which also advances the SPL development by supplementing a rational dynamic curriculum constraint. The new model finely accords with the "instructor-student-collaborative" learning mode in human education. The advantages of this proposed framework are two-folds: i) The required number of annotated samples is significantly decreased while the comparable performance is guaranteed. A dramatic reduction of user effort is also achieved over other state-of-the-art active learning techniques. ii) The mixture of SPL and AL effectively improves not only the classifier accuracy compared to existing AL/SPL methods but also the robustness against noisy data. We evaluate our framework on two challenging datasets, and demonstrate very promising results. (http://hcp.sysu.edu.cn/projects/aspl/), Comment: To appear in IEEE Transactions on Pattern Analysis and Machine Intelligence 2017

Human pose estimation (i.e., locating the body parts / joints of a person) is a fundamental problem in human-computer interaction and multimedia applications. Significant progress has been made based on the development of depth sensors, i.e., accessible human pose prediction from still depth images [32]. However, most of the existing approaches to this problem involve several components/models that are independently designed and optimized, leading to suboptimal performances. In this paper, we propose a novel inference-embedded multi-task learning framework for predicting human pose from still depth images, which is implemented with a deep architecture of neural networks. Specifically, we handle two cascaded tasks: i) generating the heat (confidence) maps of body parts via a fully convolutional network (FCN); ii) seeking the optimal configuration of body parts based on the detected body part proposals via an inference built-in MatchNet [10], which measures the appearance and geometric kinematic compatibility of body parts and embodies the dynamic programming inference as an extra network layer. These two tasks are jointly optimized. Our extensive experiments show that the proposed deep model significantly improves the accuracy of human pose estimation over other several state-of-the-art methods or SDKs. We also release a large-scale dataset for comparison, which includes 100K depth images under challenging scenarios., Comment: To appear in ACM Multimedia 2016, full paper (oral), 10 pages, 11 figures

Recently, machine learning based single image super resolution (SR) approaches focus on jointly learning representations for high-resolution (HR) and low-resolution (LR) image patch pairs to improve the quality of the super-resolved images. However, due to treat all image pixels equally without considering the salient structures, these approaches usually fail to produce visual pleasant images with sharp edges and fine details. To address this issue, in this work we present a new novel SR approach, which replaces the main building blocks of the classical interpolation pipeline by a flexible, content-adaptive deep neural networks. In particular, two well-designed structure-aware components, respectively capturing local- and holistic- image contents, are naturally incorporated into the fully-convolutional representation learning to enhance the image sharpness and naturalness. Extensively evaluations on several standard benchmarks (e.g., Set5, Set14 and BSD200) demonstrate that our approach can achieve superior results, especially on the image with salient structures, over many existing state-of-the-art SR methods under both quantitative and qualitative measures., Comment: Published on ICME 2016 (oral), 6 pages, 6 figures

Understanding human activity is very challenging even with the recently developed 3D/depth sensors. To solve this problem, this work investigates a novel deep structured model, which adaptively decomposes an activity instance into temporal parts using the convolutional neural networks (CNNs). Our model advances the traditional deep learning approaches in two aspects. First, { we incorporate latent temporal structure into the deep model, accounting for large temporal variations of diverse human activities. In particular, we utilize the latent variables to decompose the input activity into a number of temporally segmented sub-activities, and accordingly feed them into the parts (i.e. sub-networks) of the deep architecture}. Second, we incorporate a radius-margin bound as a regularization term into our deep model, which effectively improves the generalization performance for classification. For model training, we propose a principled learning algorithm that iteratively (i) discovers the optimal latent variables (i.e. the ways of activity decomposition) for all training instances, (ii) { updates the classifiers} based on the generated features, and (iii) updates the parameters of multi-layer neural networks. In the experiments, our approach is validated on several complex scenarios for human activity recognition and demonstrates superior performances over other state-of-the-art approaches., Comment: 16 pages, 9 figures, to appear in International Journal of Computer Vision 2015

Driven by recent vision and graphics applications such as image segmentation and object recognition, computing pixel-accurate saliency values to uniformly highlight foreground objects becomes increasingly important. In this paper, we propose a unified framework called PISA, which stands for Pixelwise Image Saliency Aggregating various bottom-up cues and priors. It generates spatially coherent yet detail-preserving, pixel-accurate and fine-grained saliency, and overcomes the limitations of previous methods which use homogeneous superpixel-based and color only treatment. PISA aggregates multiple saliency cues in a global context such as complementary color and structure contrast measures with their spatial priors in the image domain. The saliency confidence is further jointly modeled with a neighborhood consistence constraint into an energy minimization formulation, in which each pixel will be evaluated with multiple hypothetical saliency levels. Instead of using global discrete optimization methods, we employ the cost-volume filtering technique to solve our formulation, assigning the saliency levels smoothly while preserving the edge-aware structure details. In addition, a faster version of PISA is developed using a gradient-driven image sub-sampling strategy to greatly improve the runtime efficiency while keeping comparable detection accuracy. Extensive experiments on a number of public datasets suggest that PISA convincingly outperforms other state-of-the-art approaches. In addition, with this work we also create a new dataset containing $800$ commodity images for evaluating saliency detection. The dataset and source code of PISA can be downloaded at http://vision.sysu.edu.cn/project/PISA/, Comment: 14 pages, 14 figures, 1 table, to appear in IEEE Transactions on Image Processing

Human activity understanding with 3D/depth sensors has received increasing attention in multimedia processing and interactions. This work targets on developing a novel deep model for automatic activity recognition from RGB-D videos. We represent each human activity as an ensemble of cubic-like video segments, and learn to discover the temporal structures for a category of activities, i.e. how the activities to be decomposed in terms of classification. Our model can be regarded as a structured deep architecture, as it extends the convolutional neural networks (CNNs) by incorporating structure alternatives. Specifically, we build the network consisting of 3D convolutions and max-pooling operators over the video segments, and introduce the latent variables in each convolutional layer manipulating the activation of neurons. Our model thus advances existing approaches in two aspects: (i) it acts directly on the raw inputs (grayscale-depth data) to conduct recognition instead of relying on hand-crafted features, and (ii) the model structure can be dynamically adjusted accounting for the temporal variations of human activities, i.e. the network configuration is allowed to be partially activated during inference. For model training, we propose an EM-type optimization method that iteratively (i) discovers the latent structure by determining the decomposed actions for each training example, and (ii) learns the network parameters by using the back-propagation algorithm. Our approach is validated in challenging scenarios, and outperforms state-of-the-art methods. A large human activity database of RGB-D videos is presented in addition., Comment: This manuscript has 10 pages with 9 figures, and a preliminary version was published in ACM MM'14 conference