Text-video retrieval aims to bridge vision and language areas, which is a crucial task in multi-modal intelligence. The core idea is to learn video and textual features to quantify their semantic relevance. A common limitation in current approaches is the oversimplification of video content, where complex spatiotemporal structures are compressed into a single global representation. Consequently, these methods struggle to fully capture dynamic visual variations and discriminative appearance inside a video, further complicating cross-modal alignment. To alleviate these issues, we introduce a novel decoupling approach that independently processes appearance and motion cues, capitalizing on their complementary nature for more expressive video modeling. Specifically, we propose an appearance-motion decomposed network (AMD-Net) to decouple spatial-level appearance and temporal-level motion understanding via the discriminative appearance learning and multi-scale motion learning modules. The proposed model enjoys several merits. First, the designed discriminative appearance learning module with a Singular Value Decomposition (SVD) based prototype initialization can effectively reduce redundant appearance information, and a high-order cross-aggregation mechanism enhances prototype resilience and facilitates comprehensive video understanding. Second, the proposed multi-scale motion learning (MML) module can capture motion features at varying temporal scales, which are complementary to appearance features for accurate text-video retrieval. Extensive experiments on five standard benchmarks demonstrate that our method performs favorably against state-of-the-art methods.

To tackle the threat of fake news, the task of detecting and grounding multi-modal media manipulation (DGM4) has received increasing attention. However, most state-of-the-art methods fail to explore the fine-grained consistency within local content, usually resulting in an inadequate perception of detailed forgery and unreliable results. In this paper, we propose a novel approach named Contextual-Semantic Consistency Learning (CSCL) to enhance the fine-grained perception ability of forgery for DGM4. Two branches for image and text modalities are established, each of which contains two cascaded decoders, i.e., Contextual Consistency Decoder (CCD) and Semantic Consistency Decoder (SCD), to capture within-modality contextual consistency and across-modality semantic consistency, respectively. Both CCD and SCD adhere to the same criteria for capturing fine-grained forgery details. To be specific, each module first constructs consistency features by leveraging additional supervision from the heterogeneous information of each token pair. Then, the forgery-aware reasoning or aggregating is adopted to deeply seek forgery cues based on the consistency features. Extensive experiments on DGM4datasets prove that CSCL achieves new state-of-the-art performance, especially for the results of grounding manipulated content. Codes and weights are avaliable at https://github.com/liyih/CSCL.

Identifying multiple novel classes in an image, known as open-vocabulary multi-label recognition, is a challenging task in computer vision. Recent studies explore the transfer of powerful vision-language models such as CLIP. However, these approaches face two critical challenges: (1) The local semantics of CLIP are disrupted due to its global pre-training objectives, resulting in unreliable regional predictions. (2) The matching property between image regions and candidate labels has been neglected, relying instead on naive feature aggregation such as average pooling, which leads to spurious predictions from irrelevant regions. In this paper, we present RAM (Recover And Match), a novel framework that effectively addresses the above issues. To tackle the first problem, we propose Ladder Local Adapter (LLA) to enforce refocusing on local regions, recovering local semantics in a memory-friendly way. For the second issue, we propose Knowledge-Constrained Optimal Transport (KCOT) to suppress meaningless matching to non-GT labels by formulating the task as an optimal transport problem. As a result, RAM achieves state-of-the-art performance on various datasets from three distinct domains, and shows great potential to boost the existing methods. Code: https://github.com/EricTan7/RAM.

Pedestrian attribute recognition has achieved high accuracy by exploring the relations between image regions and attributes. However, existing methods typically adopt features directly extracted from the backbone or utilize a single structure (e.g., transformer) to explore the relations, leading to inefficient and incomplete relation mining. To overcome these limitations, this paper proposes a comprehensive relationship framework called Vision Transformer with Relation Exploration (ViT-RE) for pedestrian attribute recognition, which includes two novel modules, namely Attribute and Contextual Feature Projection (ACFP) and Relation Exploration Module (REM). In ACFP, attribute-specific features and contextual-aware features are learned individually to capture discriminative information tailored for attributes and image regions, respectively. Then, REM employs Graph Convolutional Network (GCN) Blocks and Transformer Blocks to concurrently explore attribute, contextual, and attribute-contextual relations. To enable fine-grained relation mining, a Dynamic Adjacency Module (DAM) is further proposed to construct instance-wise adjacency matrix for the GCN Block. Equipped with comprehensive relation information, ViT-RE achieves promising performance on three popular benchmarks, including PETA, RAP, and PA-100 K datasets. Moreover, ViT-RE achieves the first place in theWACV 2023 UPAR Challenge.

In this paper, we study the problem of generalizable syn-thetic image detection, aiming to detect forgery images from diverse generative methods, e.g., GANs and diffusion mod-els. Cutting-edge solutions start to explore the benefits of pre-trained models, and mainly follow the fixed paradigm of solely training an attached classifier, e.g., combining frozen CLIP-ViT with a learnable linear layer in UniFD [43]. However, our analysis shows that such a fixed paradigm is prone to yield detectors with insufficient learning regarding forgery representations. We attribute the key challenge to the lack of forgery adaptation, and present a novel forgery-aware adaptive transformer approach, namely FatFormer. Based on the pre-trained vision-language spaces of CLIP, FatFormer introduces two core designs for the adaption to build generalized forgery representations. First, motivated by the fact that both image and frequency analysis are es-sential for synthetic image detection, we develop a forgery-aware adapter to adapt image features to discern and inte-grate local forgery traces within image and frequency do-mains. Second, we find that considering the contrastive ob-jectives between adapted image features and text prompt embeddings, a previously overlooked aspect, results in a nontrivial generalization improvement. Accordingly, we in-troduce language-guided alignment to supervise the forgery adaptation with image and text prompts in FatFormer. Ex-periments show that, by coupling these two designs, our approach tuned on 4-class ProGAN data attains a remarkable detection performance, achieving an average of 98% accu-racy to unseen GANs, and surprisingly generalizes to un-seen diffusion models with 95% accuracy.

Vision-based roadside 3D object detection has attracted rising attention in autonomous driving domain, since it en-compasses inherent advantages in reducing blind spots and expanding perception range. While previous work mainly focuses on accurately estimating depth or height for 2D-to-3D mapping, ignoring the position approximation error in the voxel pooling process. Inspired by this insight, we propose a novel voxel pooling strategy to reduce such error, dubbed BEVSpread. Specifically, instead of bringing the image features contained in a frustum point to a single BEV grid, BEVSpread considers each frustum point as a source and spreads the image features to the surrounding BEV grids with adaptive weights. To achieve superior prop- agation performance, a specific weight function is designed to dynamically control the decay speed of the weights according to distance and depth. Aided by customized CUDA parallel acceleration, BEVSpread achieves comparable inference time as the original voxel pooling. Extensive experiments on two large-scale roadside benchmarks demonstrate that, as a plug-in, BEVSpread can significantly improve the performance of existing frustum-based BEV methods by a large margin of (1.12, 5.26, 3.01) AP in vehicle, pedestrian and cyclist. The source code will be made publicly available at BEVSpread.

Text-driven video editing poses significant challenges in exhibiting flicker-free visual continuity while preserving the inherent motion patterns of original videos. Existing methods operate under a paradigm where motion and appearance are intricately intertwined. This coupling leads to the network either over-fitting appearance content - failing to capture motion patterns - or focusing on motion patterns at the expense of content generalization to diverse textual scenarios. Inspired by the pivotal role of wavelet transform in dissecting video sequences, we propose CAusal Motion Enhancement tailored for Lifting text-driven video editing (CAMEL), a novel technique with two core designs. First, we introduce motion prompts, designed to summarize motion concepts from video templates through direct optimization. The optimized prompts are purposefully integrated into latent representations of diffusion models to enhance the motion fidelity of generated results. Second, to enhance motion coherence and extend the generalization of appearance content to creative textual prompts, we propose the causal motion-enhanced attention mechanism. This mechanism is implemented in tandem with a novel causal motion filter, synergistically enhancing the motion coherence of disentangled high-frequency components, and concurrently preserving the generalization of appearance content across various textual scenarios. Extensive experimental results show the superior performance of CAMEL.

Over the past few years, deep generative models have significantly evolved, enabling the synthesis of realistic content and also bringing security concerns of illegal misuse. Therefore, active protection for generative models has been proposed recently, aiming to generate samples with hidden messages for future identification while preserving the original generating performance. However, existing active protection methods are specifically designed for generative adversarial networks (GANs), restricted to handling unconditional image generation. We observe that they get limited identification performance and visual quality when handling audio-driven video generation conditioned on target audio and source input to drive video generation with consistent context, e.g., identity and movement, between frame sequences. To address this issue, we introduce a simple yet effective activeProtection framework forAudio-DrivenVideoGeneration, named PADVG. To be specific, we present a novel frame-shared embedding module in which messages to hide are first transformed into frame-shared message coefficients. Then, these coefficients are assembled with the intermediate feature maps of video generators at multiple feature levels to generate the embedded video frames. Besides, PADVG further considers two visual consistent losses: (i) intra-frame loss is utilized to keep the visual consistency with different hidden messages; (ii) inter-frame loss is used to preserve the visual consistency across different video frames. Moreover, we also propose an auxiliary denoising training strategy through perturbing the assembled features by learnable pixel-level noise to improve identification performance, while enhancing robustness against real-world disturbances. Extensive experiments demonstrate that our proposed PADVG for audio-driven video generation can effectively identify the generated videos and achieve high visual quality.

Skeletal motions have been heavily relied upon for human activity recognition (HAR). Recently, a universal vulnerability of skeleton-based HAR has been identified across a variety of classifiers and data, calling for mitigation. To this end, we propose the first black-box defense method for skeleton-based HAR to our best knowledge. Our method is featured by full Bayesian treatments of the clean data, the adversaries and the classifier, leading to (1) a new Bayesian Energy-based formulation of robust discriminative classifiers, (2) a new adversary sampling scheme based on natural motion manifolds, and (3) a new post-train Bayesian strategy for black-box defense. We name our framework Bayesian Energy-based Adversarial Training or BEAT. BEAT is straightforward but elegant, which turns vulnerable black-box classifiers into robust ones without sacrificing accuracy. It demonstrates surprising and universal effectiveness across a wide range of skeletal HAR classifiers and datasets, under various attacks. Appendix and code are available.

In this paper, we study the problem of end-to-end multi-person pose estimation. State-of-the-art solutions adopt the DETR-like framework, and mainly develop the complex decoder, e.g., regarding pose estimation as keypoint box detection and combining with human detection in ED-Pose [38], hierarchically predicting with pose decoder and joint (keypoint) decoder in PETR [27].We present a simple yet effective transformer approach, named Group Pose. We simply regard K-keypoint pose estimation as predicting a set of N × K keypoint positions, each from a keypoint query, as well as representing each pose with an instance query for scoring N pose predictions.Motivated by the intuition that the interaction, among across-instance queries of different types, is not directly helpful, we make a simple modification to decoder self-attention. We replace single self-attention over all the N × (K + 1) queries with two subsequent group self-attentions: (i) N within-instance self-attention, with each over K keypoint queries and one instance query, and (ii) (K +1) same-type across-instance self-attention, each over N queries of the same type. The resulting decoder removes the interaction among across-instance type-different queries, easing the optimization and thus improving the performance. Experimental results on MS COCO and Crowd-Pose show that our approach without human box supervision is superior to previous methods with complex decoders, and even is slightly better than ED-Pose that uses human box supervision. Paddle1and PyTorch2codes are available.

Prompt learning has become one of the most efficient paradigms for adapting large pre-trained vision-language models to downstream tasks. Current state-of-the-art methods, like CoOp and ProDA, tend to adopt soft prompts to learn an appropriate prompt for each specific task. Recent CoCoOp further boosts the base-to-new generalization performance via an image-conditional prompt. However, it directly fuses identical image semantics to prompts of different labels and significantly weakens the discrimination among different classes as shown in our experiments. Motivated by this observation, we first propose a class-aware text prompt (CTP) to enrich generated prompts with label-related image information. Unlike CoCoOp, CTP can effectively involve image semantics and avoid introducing extra ambiguities into different prompts. On the other hand, instead of reserving the complete image representations, we propose text-guided feature tuning (TFT) to make the image branch attend to class-related representation. A contrastive loss is employed to align such augmented text and image representations on downstream tasks. In this way, the image-to-text CTP and text-to-image TFT can be mutually promoted to enhance the adaptation of VLMs for downstream tasks. Extensive experiments demonstrate that our method outperforms the existing methods by a significant margin. Especially, compared to CoCoOp, we achieve an average improvement of 4.03% on new classes and 3.19% on harmonic-mean over eleven classification benchmarks.

Visible-Infrared person re-identification is challenging due to the large modality gap. To bridge the gap, most studies heavily rely on the correlation of visible-infrared holistic person images, which may perform poorly under severe distribution shifts. In contrast, we find that some cross-modal correlated high-frequency components contain discriminative visual patterns and are less affected by variations such as wavelength, pose, and background clutter than holistic images. Therefore, we are motivated to bridge the modality gap based on such high-frequency components, and propose Prototype-guided High-frequency Patch Enhancement (ProtoHPE) with two core designs. First, to enhance the representation ability of cross-modal correlated high-frequency components, we split patches with such components by Wavelet Transform and exponential moving average Vision Transformer (ViT), then empower ViT to take the split patches as auxiliary input. Second, to obtain semantically compact and discriminative high-frequency representations of the same identity, we propose Multimodal Prototypical Contrast. To be specific, it hierarchically captures comprehensive semantics of different modal instances, facilitating the aggregation of high-frequency representations belonging to the same identity. With it, ViT can capture key high-frequency components during inference without relying on ProtoHPE, thus bringing no extra complexity. Extensive experiments validate the effectiveness of ProtoHPE.

Facial Expression Recognition (FER) in the wild is an extremely challenging task. Recently, some Vision Transformers (ViT) have been explored for FER, but most of them perform inferiorly compared to Convolutional Neural Networks (CNN). This is mainly because the new proposed modules are difficult to converge well from scratch due to lacking inductive bias and easy to focus on the occlusion and noisy areas. TransFER, a representative transformer-based method for FER, alleviates this with multi-branch attention dropping but brings excessive computations. On the contrary, we present two attentive pooling (AP) modules to pool noisy features directly. The AP modules include Attentive Patch Pooling (APP) and Attentive Token Pooling (ATP). They aim to guide the model to emphasize the most discriminative features while reducing the impacts of less relevant features. The proposed APP is employed to select the most informative patches on CNN features, and ATP discards unimportant tokens in ViT. Being simple to implement and without learnable parameters, the APP and ATP intuitively reduce the computational cost while boosting the performance by ONLY pursuing the most discriminative features. Qualitative results demonstrate the motivations and effectiveness of our attentive poolings. Besides, quantitative results on six in-the-wild datasets outperform other state-of-the-art methods.

Facial age estimation has attracted considerable attention owing to its great potential in applications. However, it still falls short of reliable age estimation due to the lack of sufficient training data with accurate age labels. Using conventional semi-supervised methods to exploit unlabeled data appears to be a good solution, but it does not yield sufficient performance gains while significantly increasing training time. Therefore, to tackle these problems, we present a Divergence-driven Consistency Training (DCT) method for enhancing both efficiency and performance in this paper. Following the idea of pseudo-labeling and consistency regularization, we assign pseudo labels predicted by the teacher model to unlabeled samples and then train the student model on labeled and unlabeled samples based on consistency regularization. Based on this, we propose two main promotions. The first is the Efficient Sample Selection (ESS) strategy, which is based on the Divergence Score to select effective samples from massive unlabeled images to reduce the training time and improve efficiency. The second is Identity Consistency (IC) regularization as the additional loss function, which introduces a high dependency of aging traits on a person. Moreover, we propose Local Prediction (LP), which is a plug-and-play component, to capture local semantics. Extensive experiments on multiple age benchmark datasets, including CACD, Morph II, MIVIA, and Chalearn LAP 2015, indicate DCT outperforms the state-of-the-art approaches significantly.

The availability of handy multi-modal (i.e., RGB-D) sensors has brought about a surge of face anti-spoofing research. However, the current multi-modal face presentation attack detection (PAD) has two defects: (1) The framework based on multi-modal fusion requires providing modalities consistent with the training input, which seriously limits the deployment scenario. (2) The performance of ConvNet-based model on high fidelity datasets is increasingly limited. In this work, we present a pure transformer-based framework, dubbed the Flexible Modal Vision Transformer (FM-ViT), for face anti-spoofing to flexibly target any single-modal (i.e., RGB) attack scenarios with the help of available multi-modal data. Specifically, FM-ViT retains a specific branch for each modality to capture different modal information and introduces the Cross-Modal Transformer Block (CMTB), which consists of two cascaded attentions named Multi-headed Mutual-Attention (MMA) and Fusion-Attention (MFA) to guide each modal branch to mine potential features from informative patch tokens, and to learn modality-agnostic liveness features by enriching the modal information of own CLS token, respectively. Experiments demonstrate that the single model trained based on FM-ViT can not only flexibly evaluate different modal samples, but also outperforms existing single-modal frameworks by a large margin, and approaches the multi-modal frameworks introduced with smaller FLOPs and model parameters.

In recent years, face forgery detectors have aroused great interest and achieved impressive performance, but they are still struggling with generalization and robustness. In this work, we explore taking full advantage of the fine-grained forgery traces in both spatial and frequency domains to alleviate this issue. Specifically, we propose a novel High-Frequency Fine-Grained Transformer (F2Trans) network which contains two important components, namely Central Difference Attention (CDA) and High-frequency Wavelet Sampler (HWS). The premier CDA module is capable of capturing invariant fine-grained manipulation patterns by aggregating both pixel-level intensity and gradient information of the query to generate key and value pairs. Subsequently, the proposed HWS discards the low-frequency components of wavelet transformation and hierarchically explores high-frequency forgery cues of feature maps, which prevents model confusion caused by low-frequency components and pays attention to local frequency information. In addition, HWS can be employed as a special pooling layer for the F2Trans architecture to produce hierarchical feature representations in the spatial-frequency domain. Extensive experiments on multiple popular benchmarks demonstrate the generalization and robustness of the specially designed F2Trans framework is well-tailored for face forgery detection when confronting the cross-dataset, cross-manipulation, and unseen perturbations.

Facial age estimation has received a lot of attention for its diverse application scenarios. Most existing studies treat each sample equally and aim to reduce the average estimation error for the entire dataset, which can be summarized as General Age Estimation. However, due to the long-tailed distribution prevalent in the dataset, treating all samples equally will inevitably bias the model toward the head classes (usually the adult with a majority of samples). Driven by this, some works suggest that each class should be treated equally to improve performance in tail classes (with a minority of samples), which can be summarized as Long-tailed Age Estimation. However, Long-tailed Age Estimation usually faces a performance trade-off, i.e., achieving improvement in tail classes by sacrificing the head classes. In this paper, our goal is to design a unified framework to perform well on both tasks, killing two birds with one stone. To this end, we propose a simple, effective, and flexible training paradigm named GLAE, which is two-fold. First, we propose Feature Rearrangement (FR) and Pixel-level Auxiliary learning (PA) for better feature utilization to improve the overall age estimation performance. Second, we propose Adaptive Routing (AR) for selecting the appropriate classifier to improve performance in the tail classes while maintaining the head classes. Moreover, we introduce a new metric, named Class-wise Mean Absolute Error (CMAE), to equally evaluate the performance of all classes. Our GLAE provides a surprising improvement on Morph II, reaching the lowest MAE and CMAE of 1.14 and 1.27 years, respectively. Compared to the previous best method, MAE dropped by up to 34%, which is an unprecedented improvement, and for the first time, MAE is close to 1 year old. Extensive experiments on other age benchmark datasets, including CACD, MIVIA, and Chalearn LAP 2015, also indicate that GLAE outperforms the state-of-the-art approaches significantly.

The networks trained on the long-tailed dataset vary remarkably, despite the same training settings, which shows the great uncertainty in long-tailed learning. To alleviate the uncertainty, we propose a Nested Collaborative Learning (NCL), which tackles the problem by collaboratively learning multiple experts together. NCL consists of two core components, namely Nested Individual Learning (NIL) and Nested Balanced Online Distillation (NBOD), which focus on the individual supervised learning for each single expert and the knowledge transferring among multiple experts, respectively. To learn representations more thoroughly, both NIL and NBOD are formulated in a nested way, in which the learning is conducted on not just all categories from a full perspective but some hard categories from a partial perspective. Regarding the learning in the partial perspective, we specifically select the negative categories with high predicted scores as the hard categories by using a proposed Hard Category Mining (HCM). In the NCL, the learning from two perspectives is nested, highly related and complementary, and helps the network to capture not only global and robust features but also meticulous distinguishing ability. Moreover, self-supervision is further utilized for feature enhancement. Extensive experiments manifest the superiority of our method with outperforming the state-of-the-art whether by using a single model or an ensemble. Code is available at https://github.com/Bazinga699/NCL

Deepfake detection has attracted increasing attention in recent years. In this paper, we propose a transformer-based framework with feature compensation and aggregation (Trans-FCA) to extract rich forgery cues for deepfake detection. To compensate local features for transformers, we propose a Locality Compensation Block (LCB) containing a Global-Local Cross-Attention (GLCA) to attentively fuse global transformer features and local convolutional features. To aggregate features of all layers for capturing comprehensive and various fake flaws, we propose Multi-head Clustering Projection (MCP) and Frequency-guided Fusion Module (FFM), where the MCP attentively reduces redundant features into a few concentrated clusters, and the FFM interacts all clustered features under the guidance of frequency cues. In Trans-FCA, besides global cues captured by transformer architecture, local details and rich forgery defects are also captured using the proposed fetaure compensation and aggregation. Extensive experiments show our method outperforms the state-of-the-art methods on both intra-dataset and cross-dataset testings (with AUCs of 99.85% on FaceForensics++ and 78.57% on Celeb-DF), which clearly demonstrates the superiority of our Trans-FCA for deepfake detection.

Face presentation attack detection (PAD) is essential to secure face recognition systems primarily from high-fidelity mask attacks. Most existing 3D mask PAD benchmarks suffer from several drawbacks: 1) a limited number of mask identities, types of sensors, and a total number of videos; 2) low-fidelity quality of facial masks. Basic deep models and remote photoplethysmography (rPPG) methods achieved acceptable performance on these benchmarks but still far from the needs of practical scenarios. To bridge the gap to real-world applications, we introduce a large-scale High-Fidelity Mask dataset, namely HiFiMask. Specifically, a total amount of 54,600 videos are recorded from 75 subjects with 225 realistic masks by 7 new kinds of sensors. Along with the dataset, we propose a novel Contrastive Context-aware Learning (CCL) framework. CCL is a new training methodology for supervised PAD tasks, which is able to learn by leveraging rich contexts accurately (e.g., subjects, mask material and lighting) among pairs of live faces and high-fidelity mask attacks. Extensive experimental evaluations on HiFiMask and three additional 3D mask datasets demonstrate the effectiveness of our method. The codes and dataset will be released soon.

Recently, face manipulation techniques have caused increasing trust concerns in our society. Although current face manipulation detection methods achieve impressive performance regarding intra-dataset evaluation, they are struggling to improve the generalization and robustness ability. To address this issue, we propose a novel Hierarchical Frequency-assisted Interactive Networks (HFI-Net) to explore comprehensive frequency-related forgery cues for face manipulation detection. At first, we formulate HFI-Net as a dual-branch network to take full advantage of both CNN and transformer for capturing local details and global context information, respectively. Considering the forged faces are easy to show flaws in the frequency domain, a novel Frequency-based Feature Refinement (FFR) module is proposed to learn frequency-based attention from RGB features. FFR module emphasizes forgery cues and suppresses the pristine semantics information by keeping middle-high frequency features while discarding the low-frequency ones. Based on FFR, we further develop a co-sharing Global-Local Interaction (GLI) module to conduct frequency-assisted interactions while capturing complementarity among dual branches. Lastly, we further implement the GLI module in each stage of the network to effectively explore multi-level frequency artifacts. Extensive experiments are conducted on several popular benchmarks including FaceForensics++, Celeb-DF, DeepFake-TIMIT, DFDC, UADFV, and DeeperForensics-1.0, which shows that our model outperforms the state-of-the-art, especially in unseen datasets, manipulations, and perturbations evaluation.

In our daily life, a large number of activities require identity verification, e.g., ePassport gates. Most of those verification systems recognize who you are by matching the ID document photo (ID face) to your live face image (spot face). The ID vs. Spot (IvS) face recognition is different from general face recognition where each dataset usually contains a small number of subjects and sufficient images for each subject. In IvS face recognition, the datasets usually contain massive class numbers (million or more) while each class only has two image samples (one ID face and one spot face), which makes it very challenging to train an effective model (e.g., excessive demand on GPU memory if conducting the classification on such massive classes, hardly capture the effective features for bisample data of each identity, etc.). To avoid the excessive demand on GPU memory, a two-stage training method is developed, where we first train the model on the dataset in general face recognition (e.g., MS-Celeb-1M) and then employ the metric learning losses (e.g., triplet and quadruplet losses) to learn the features on IvS data with million classes. To extract more effective features for IvS face recognition, we propose two novel algorithms to enhance the network by selecting harder samples for training. Firstly, a Cross-Batch Hard Example Mining (CB-HEM) is proposed to select the hard triplets from not only the current mini-batch but also past dozens of mini-batches (for convenience, we use batch to denote a mini-batch in the following), which can significantly expand the space of sample selection. Secondly, a Pseudo Large Batch (PLB) is proposed to virtually increase the batch size with a fixed GPU memory. The proposed PLB and CB-HEM can be employed simultaneously to train the network, which dramatically expands the selecting space by hundreds of times, where the very hard sample pairs especially the hard negative pairs can be selected for training to enhance the discriminative capability. Extensive comparative evaluations conducted on multiple IvS benchmarks demonstrate the effectiveness of the proposed method.

The issue of ethnic bias has proven to affect the performance of face recognition in previous works, while it still remains to be vacant in face anti-spoofing. Therefore, in order to study the ethnic bias for face anti-spoofing, we introduce the largest CASIA-SURF Cross-ethnicity Face Anti-spoofing (CeFA) dataset, covering 3 ethnicities, 3 modalities, 1,607 subjects, and 2D plus 3D attack types. Five protocols are introduced to measure the affect under varied evaluation conditions, such as cross-ethnicity, unknown spoofs or both of them. As our knowledge, CASIA-SURF CeFA is the first dataset including explicit ethnic labels in current released datasets. Then, we propose a novel multi-modal fusion method as a strong baseline to alleviate the ethnic bias, which employs a partially shared fusion strategy to learn complementary information from multiple modalities. Extensive experiments have been conducted on the proposed dataset to verify its significance and generalization capability for other existing datasets, i.e., CASIA-SURF, OULU-NPU and SiW datasets. The dataset is available at https://sites.google.com/qq.com/face-anti-spoofing/welcome/challengecvpr2020?authuser=0.

Face Presentation Attack Detection (PAD) approaches based on multi-modal data have been attracted increasingly by the research community. However, they require multi-modal face data consistently involved in both the training and testing phases. It would severely limit the applicability due to the most Face Anti-spoofing (FAS) systems are only equipped with Visible (VIS) imaging devices, i.e., RGB cameras. Therefore, how to use other modality (i.e., Near-Infrared (NIR)) to assist the performance improvement of VIS-based PAD is significant for FAS. In this work, we first discuss the big gap of performances among different modalities even though the same backbone network is applied. Then, we propose a novel Cross-modal Auxiliary (CMA) framework for the VIS-based FAS task. The main trait of CMA is that the performance can be greatly improved with the help of other modality while no other modality is required in the testing stage. The proposed CMA consists of a Modality Translation Network (MT-Net) and a Modality Assistance Network (MA-Net). The former aims to close the visible gap between different modalities via a generative model that maps inputs from one modality (i.e., RGB) to another (i.e., NIR). The latter focuses on how to use the translated modality (i.e., target modality) and RGB modality (i.e., source modality) together to train a discriminative PAD model. Extensive experiments are conducted to demonstrate that the proposed framework can push the state-of-the-art (SOTA) performances on both multi-modal datasets (i.e., CASIA-SURF, CeFA, and WMCA) and RGB-based datasets (i.e., OULU-NPU, and SiW).

In this paper, we propose a new end-to-end network, named Joint Learning of Attribute and Contextual relations (JLAC), to solve the task of pedestrian attribute recognition. It includes two novel modules: Attribute Relation Module (ARM) and Contextual Relation Module (CRM). For ARM, we construct an attribute graph with attribute-specific features which are learned by the constrained losses, and further use Graph Convolutional Network (GCN) to explore the correlations among multiple attributes. For CRM, we first propose a graph projection scheme to project the 2-D feature map into a set of nodes from different image regions, and then employ GCN to explore the contextual relations among those regions. Since the relation information in the above two modules is correlated and complementary, we incorporate them into a unified framework to learn both together. Experiments on three benchmarks, including PA-100K, RAP, PETA attribute datasets, demonstrate the effectiveness of the proposed JLAC.

Facial expression recognition is inherently a challenging task, especially for the in-the-wild images with various occlusions and large pose variations, which may lead to the loss of some crucial information. To address it, in this paper, we propose an attentive hybrid architecture (AHA) which learns global, local and integrated features based on different face regions. Compared with one type of feature, our extracted features own complementary information and can reduce the loss of crucial information. Specifically, AHA contains three branches, where all sub-networks in those branches employ the attention mechanism to further localize the interested pixels/regions. Moreover, we propose a two-step fusion strategy based on LSTM to deeply explore the hidden correlations among different face regions. Extensive experiments on four popular expression databases (i.e., CK+, FER-2013, SFEW 2.0, RAF-DB) show the effectiveness of the proposed method.

In this paper, we propose a novel unified network named Deep Hybrid-Aligned Architecture for facial age estimation. It contains global, local and global-local branches. They are jointly optimized and thus can capture multiple types of features with complementary information. In each branch, we employ a separate loss for each sub-network to extract the independent features and use a recurrent fusion to explore correlations among those region features. Considering that the pose variations may lead to misalignment in different regions, we design an Aligned Region Pooling operation to generate aligned region features. Moreover, a new large age dataset named Web-FaceAge owning more than 120K samples is collected under diverse scenes and spanning a large age range. Experiments on five age benchmark datasets, including Web-FaceAge, Morph, FG-NET, CACD and Chalearn LAP 2015, show that the proposed method outperforms the state-of-the-art approaches significantly.

Recognizing the pedestrian attributes in surveillance scenes is an inherently challenging task, especially for the pedestrian images with large pose variations, complex backgrounds, and various camera viewing angles. To select important and discriminative regions or pixels against the variations, three attention mechanisms are proposed, including parsing attention, label attention, and spatial attention. Those attentions aim at accessing effective information by considering problems from different perspectives. To be specific, the parsing attention extracts discriminative features by learning not only where to turn attention to but also how to aggregate features from different semantic regions of human bodies, e.g., head and upper body. The label attention aims at targetedly collecting the discriminative features for each attribute. Different from the parsing and label attention mechanisms, the spatial attention considers the problem from a global perspective, aiming at selecting several important and discriminative image regions or pixels for all attributes. Then, we propose a joint learning framework formulated in a multi-task-like way with these three attention mechanisms learned concurrently to extract complementary and correlated features. This joint learning framework is named Joint Learning of Parsing attention, Label attention, and Spatial attention for Pedestrian Attributes Analysis (JLPLS-PAA, for short). Extensive comparative evaluations conducted on multiple large-scale benchmarks, including PA-100K, RAP, PETA, Market-1501, and Duke attribute datasets, further demonstrate the effectiveness of the proposed JLPLS-PAA framework for pedestrian attribute analysis.

Different ages are closely related especially among the adjacent ages because aging is a slow and extremely non-stationary process with much randomness. To explore the relationship between the real age and its adjacent ages, an age group-n encoding (AGEn) method is proposed in this paper. In our model, adjacent ages are grouped into the same group and each age corresponds to n groups. The ages grouped into the same group would be regarded as an independent class in the training stage. On this basis, the original age estimation problem can be transformed into a series of binary classification sub-problems. And a deep Convolutional Neural Networks (CNN) with multiple classifiers is designed to cope with such sub-problems. Later, a Local Age Decoding (LAD) strategy is further presented to accelerate the prediction process, which locally decodes the estimated age value from ordinal classifiers. Besides, to alleviate the imbalance data learning problem of each classifier, a penalty factor is inserted into the unified objective function to favor the minority class. To compare with state-of-the-art methods, we evaluate the proposed method on FG-NET, MORPH II, CACD and Chalearn LAP 2015 databases and it achieves the best performance.

Owing to the variations including both intrinsic and extrinsic factors, age estimation remains a challenging problem. In this paper, five cascaded structure frameworks are proposed for age estimation based on convolutional neural networks. All frameworks are learned and guided by auxiliary demographic information, since other demographic information (i.e., gender and race) is beneficial for age prediction. Each cascaded structure framework is embodied in a parent network and several subnetworks. For example, one of the applied framework is a gender classifier trained by gender information, and then two subnetworks are trained by the male and female samples, respectively. Furthermore, we use the features extracted from the cascaded structure frameworks with Gaussian process regression that can boost the performance further for age estimation. Experimental results on the MORPH II and CACD datasets have gained superior performances compared to the state-of-the-art methods. The mean absolute error is significantly reduced from 3.63 to 2.93 years under the same test protocol on the MORPH II dataset.