IEEE transactions on biometrics, behavior, and identity science最新文献_第3页

Revocable Biometric System Using Multiple Identical Deep Neural Matcher Instances 使用多个相同深度神经匹配器实例的可撤销生物识别系统

IF 4.4

IEEE transactions on biometrics, behavior, and identity science Pub Date : 2026-09-01 Epub Date: 2026-03-09 DOI: 10.1109/TBIOM.2026.3671674

Aman Bhatta;Genesis Argueta;Michael C. King;Kevin W. Bowyer

{"title":"Revocable Biometric System Using Multiple Identical Deep Neural Matcher Instances","authors":"Aman Bhatta;Genesis Argueta;Michael C. King;Kevin W. Bowyer","doi":"10.1109/TBIOM.2026.3671674","DOIUrl":"https://doi.org/10.1109/TBIOM.2026.3671674","url":null,"abstract":"Biometric authentication systems face a critical security challenge: once a biometric template is compromised, it cannot be replaced. Revocable biometrics address this by enabling template cancellation and reissuance without degrading recognition accuracy. We systematically evaluate a revocable framework in which multiple independently trained instances of the same end-to-end model yield revocable templates. From a shared backbone architecture, training dataset, and identical hyperparameter configurations, we derive distinct face matchers that exhibit near-identical genuine and impostor score distributions. Notably, cross-model similarity scores (genuine scores) for the same individual are lower than same-model scores for different individuals (impostor scores), reducing the value of stolen templates to sub-impostor levels in cross-model scenarios. We assess ResNet-100, MobileFaceNet, and ViT backbones under ArcFace and AdaFace losses, verifying robustness across architectures and loss functions. t-SNE visualizations of embeddings from independently trained instances with the same backbone and loss show that features realize different coordinate systems on a shared face-identity manifold, achieving comparable validity and accuracy rather than representing trivial permutations or transforms. However, an adversary with paired feature vectors from two models for the same enrolled images can still approximate the inter-model transformation. To mitigate this, we propose a model-mixture deployment that randomly and uniformly partitions subjects across models, with subsequent updates following the same randomized partitioning to obstruct direct linear mappings while maintaining stable operation of the revocable biometric system.","PeriodicalId":73307,"journal":{"name":"IEEE transactions on biometrics, behavior, and identity science","volume":"8 5","pages":"792-804"},"PeriodicalIF":4.4,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148961949","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

FAS-MBCA: Face Anti-Spoofing Using Multi-Path Bi-Directional Cross-Attention Fusion FAS-MBCA：基于多路径双向交叉注意融合的人脸防欺骗技术

IF 4.4

IEEE transactions on biometrics, behavior, and identity science Pub Date : 2026-09-01 Epub Date: 2026-03-12 DOI: 10.1109/TBIOM.2026.3692509

Dhimas Arief Dharmawan;Made Vidyatma Adhi Krisna;Anto Satriyo Nugroho;Andrey Ferriyan;Bambang Yuwono;Jaconiah Widiwasita;Aqsha Jauzaarafa' Setya Hadi

{"title":"FAS-MBCA: Face Anti-Spoofing Using Multi-Path Bi-Directional Cross-Attention Fusion","authors":"Dhimas Arief Dharmawan;Made Vidyatma Adhi Krisna;Anto Satriyo Nugroho;Andrey Ferriyan;Bambang Yuwono;Jaconiah Widiwasita;Aqsha Jauzaarafa' Setya Hadi","doi":"10.1109/TBIOM.2026.3692509","DOIUrl":"https://doi.org/10.1109/TBIOM.2026.3692509","url":null,"abstract":"Face Anti-Spoofing (FAS) is a critical component for securing face recognition systems against diverse presentation attacks. multi-modal FAS approaches leverage complementary cues from multiple sensors; however, existing methods often adopt single-path fusion or unidirectional interaction schemes, which may introduce modality hierarchy and limit comprehensive cross-modal feature refinement. In this paper, we propose FAS-MBCA, a structured multi-path bi-directional cross-attention framework for tri-modal face anti-spoofing. FAS-MBCA explicitly constructs a complete pairwise interaction topology among RGB, depth, and infrared modalities, enabling reciprocal feature-map refinement across all modality pairs. By modeling six-direction cross-modal interactions within a lightweight CNN-based architecture, the proposed framework enhances multi-modal synergy while maintaining computational efficiency. We evaluate FAS-MBCA on multiple widely used benchmarks, including CASIA-SURF (MmFA), CASIA-CeFA, and WMCA, under both intra-dataset and cross-dataset protocols. Extensive experiments demonstrate that FAS-MBCA consistently outperforms single-modal baselines and naïve fusion strategies, and achieves competitive or superior performance compared to recent multi-modal methods. Efficiency analysis in terms of model parameters and GFLOPs further indicates a favorable accuracy–efficiency trade-off. These results validate that explicitly modeling complete bi-directional cross-modal interaction at the feature-map level provides an effective and practical solution for multi-modal face anti-spoofing. The source code of FAS-MBCA is publicly available at <uri>https://github.com/nascafas2025/FAS-MBCA</uri>","PeriodicalId":73307,"journal":{"name":"IEEE transactions on biometrics, behavior, and identity science","volume":"8 5","pages":"947-959"},"PeriodicalIF":4.4,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148961805","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

IEEE Transactions on Biometrics, Behavior, and Identity Science Publication Information IEEE生物计量学、行为与身份科学学报

IF 4.4

IEEE transactions on biometrics, behavior, and identity science Pub Date : 2026-09-01 Epub Date: 2026-09-14 DOI: 10.1109/TBIOM.2026.3719457

引用次数: 0

Bidirectional Feature Enhancement for Video-Based Occluded Cloth-Changing Person Re-Identification 基于视频遮挡换衣人再识别的双向特征增强

IF 4.4

IEEE transactions on biometrics, behavior, and identity science Pub Date : 2026-09-01 Epub Date: 2026-03-06 DOI: 10.1109/TBIOM.2026.3690988

Vuong D. Nguyen;Pranav Mantini;Shishir K. Shah

{"title":"Bidirectional Feature Enhancement for Video-Based Occluded Cloth-Changing Person Re-Identification","authors":"Vuong D. Nguyen;Pranav Mantini;Shishir K. Shah","doi":"10.1109/TBIOM.2026.3690988","DOIUrl":"https://doi.org/10.1109/TBIOM.2026.3690988","url":null,"abstract":"Video-based Person Re-Identification (Re-ID) has seen significant progress with deep learning techniques aimed at extracting robust spatial features from frame sequences. However, existing methods suffer from key limitations, particularly in dealing with occlusions and clothing changes, which are prevalent in real-world scenarios. In this work, we address the task termed Video-based Occluded Cloth-Changing Re-ID (VOCCRe-ID), which aims to address the compounded challenges of occlusions and clothing changes in video sequences. We propose Bidirectional Feature Enhancement (BiFE), a method that uses a novel occlusion synthesis strategy, synthesizing occlusions only on middle consecutive frames and allowing their positions to shift dynamically, thereby mimicking more realistic occlusion scenarios. Our method utilizes non-occluded keyframes at both ends of the sequence for bidirectional feature propagation, ensuring temporal consistency and filling occlusion gaps. We also introduce an attention mechanism to correct noise and ambiguity in the propagated features, which are then used to enhance both appearance and shape features. To provide a more practical benchmark, we rebuild the Occluded-VCCR dataset using the proposed occlusion synthesis strategy, making it better approximate real-world occlusion patterns, despite being synthetically generated. Experimental results demonstrate significant improvements in VOCCRe-ID performance, showing that our approach effectively handles occlusions and clothing changes while ensuring robust feature representation.","PeriodicalId":73307,"journal":{"name":"IEEE transactions on biometrics, behavior, and identity science","volume":"8 5","pages":"741-753"},"PeriodicalIF":4.4,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148961888","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

IEEE Transactions on Biometrics, Behavior, and Identity Science Information for Authors IEEE生物识别、行为和身份科学信息作者汇刊

IF 4.4

IEEE transactions on biometrics, behavior, and identity science Pub Date : 2026-09-01 Epub Date: 2026-09-14 DOI: 10.1109/TBIOM.2026.3719458

引用次数: 0

On the Illusion of Gender Bias in Face Recognition: Explaining the Fairness Issue Through Non-Demographic Attributes 论人脸识别中的性别偏见错觉：用非人口统计学属性解释公平问题

IF 4.4

IEEE transactions on biometrics, behavior, and identity science Pub Date : 2026-09-01 Epub Date: 2026-03-29 DOI: 10.1109/TBIOM.2026.3698181

Paul Jonas Kurz;Haiyu Wu;Rouqaiah Al-Refai;Kevin W. Bowyer;Philipp Terhörst

{"title":"On the Illusion of Gender Bias in Face Recognition: Explaining the Fairness Issue Through Non-Demographic Attributes","authors":"Paul Jonas Kurz;Haiyu Wu;Rouqaiah Al-Refai;Kevin W. Bowyer;Philipp Terhörst","doi":"10.1109/TBIOM.2026.3698181","DOIUrl":"https://doi.org/10.1109/TBIOM.2026.3698181","url":null,"abstract":"Face recognition systems (FRS) exhibit significant accuracy differences based on the user’s gender. Since such a gender gap reduces the trustworthiness of FRS, more recent efforts have tried to find the causes. However, these studies make use of manually selected, correlated, and small-sized sets of facial features to support their claims. In this work, we analyze gender bias in face recognition by successfully extending the search domain to decorrelated combinations of 40 non-demographic facial characteristics. First, we introduce a toolchain to effectively decorrelate and aggregate facial attributes to enable a less-biased gender analysis on large-scale data. Second, we tailor two specialized metrics to quantify the effect of facial attributes on absolute and relative fairness. Based on these grounds, we thirdly present a novel unsupervised joint investigation framework capable of identifying attribute combinations leading to vanishing bias when used as filter predicates for balanced testing datasets. Experiments show the gender gap vanishing when images of male and female subjects share specific attributes, clearly indicating that the disparate performance is not a question of biology but of the social definition of appearance. These findings could reshape our understanding of fairness in face biometrics and provide insights into FRS, helping to address gender bias issues.","PeriodicalId":73307,"journal":{"name":"IEEE transactions on biometrics, behavior, and identity science","volume":"8 5","pages":"931-946"},"PeriodicalIF":4.4,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148961834","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

IEEE Transactions on Biometrics, Behavior, and Identity Science Information for Authors IEEE生物识别、行为和身份科学信息作者汇刊

IF 4.4

IEEE transactions on biometrics, behavior, and identity science Pub Date : 2026-07-01 Epub Date: 2026-06-17 DOI: 10.1109/TBIOM.2026.3688419

引用次数: 0

Structural Prior-Guided Cross-Attention Network for OCT Fingerprint Presentation Attack Detection 基于结构先验引导交叉注意网络的OCT指纹表示攻击检测

IF 4.4

IEEE transactions on biometrics, behavior, and identity science Pub Date : 2026-07-01 Epub Date: 2026-03-27 DOI: 10.1109/TBIOM.2026.3678255

Yi-Peng Liu;Jiajin Qi;Zhanqing Li;Jie Ji;Haohao Sun;Yilong Zhang;Haixia Wang;Jing Li

{"title":"Structural Prior-Guided Cross-Attention Network for OCT Fingerprint Presentation Attack Detection","authors":"Yi-Peng Liu;Jiajin Qi;Zhanqing Li;Jie Ji;Haohao Sun;Yilong Zhang;Haixia Wang;Jing Li","doi":"10.1109/TBIOM.2026.3678255","DOIUrl":"https://doi.org/10.1109/TBIOM.2026.3678255","url":null,"abstract":"Fingerprint presentation attack detection (PAD) is crucial in automatic fingerprint recognition systems. Optical coherence tomography (OCT) can capture the 3D complex structure of a fingertip because of its high resolution and skin penetrability, thus providing more reliable PAD performance for fingerprint recognition technology. Unsupervised OCT fingerprint PAD remains an unresolved issue. Existing methods rely primarily on whole slices of OCT volume and often overlook the importance of continuity and correlation among multiple features beneath the skin. In this study, we explore the reconstruction and segmentation tasks for OCT B-scan slices to discover the potential relationships between the multilayer tissue structure and overall features, and their implications for PAD tasks. We propose a novel cross-attention network that leverages prior knowledge of fingertip tissue structure. This network enhances the PAD performance by incorporating a cross-attention segmentation task interaction module to capture more structural information. Notably, our method does not require presentation attack (PA) samples during training and exhibits good generalizability for PA samples made from various materials during testing. Moreover, we introduce a reconstruction loss, which is used to further widen the reconstruction distance between the bona fide and PA samples to obtain a more accurate decision boundary for the PAD task. The source code will be made publicly available at <uri>https://github.com/SoberDrunkQ/SPCA-Net</uri>","PeriodicalId":73307,"journal":{"name":"IEEE transactions on biometrics, behavior, and identity science","volume":"8 4","pages":"612-625"},"PeriodicalIF":4.4,"publicationDate":"2026-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148433349","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A Survey on Deep Learning for Face Anti-Spoofing in Automotive Applications 深度学习在汽车人脸防欺骗中的应用研究

IF 4.4

IEEE transactions on biometrics, behavior, and identity science Pub Date : 2026-07-01 Epub Date: 2026-02-27 DOI: 10.1109/TBIOM.2026.3669032

Michael Ruiz;Hossein Hassani;Soodeh Nikan

{"title":"A Survey on Deep Learning for Face Anti-Spoofing in Automotive Applications","authors":"Michael Ruiz;Hossein Hassani;Soodeh Nikan","doi":"10.1109/TBIOM.2026.3669032","DOIUrl":"https://doi.org/10.1109/TBIOM.2026.3669032","url":null,"abstract":"Facial recognition (FR) is increasingly deployed in vehicles for authentication and personalized user settings. However, these systems are vulnerable to sophisticated spoofing attacks, necessitating robust face anti-spoofing (FAS) solutions. Deep learning (DL)-based FAS has achieved strong results in general settings, but its adaptation to in-vehicle environments faces distinct challenges, including dynamic lighting, frequent occlusions, and limited onboard computational resources. Although there is extensive research on general FAS, its application and performance within in-vehicle environments remain underexplored. This paper reviews the state of the art DL-based FAS methods and evaluates their applicability to automotive contexts. The survey covers a detailed taxonomy of traditional and advanced approaches, including Convolutional Neural Networks (CNNs), Vision Transformers (ViT), generative models, multimodal and spatiotemporal frameworks, adversarially robust techniques, and data-efficient paradigms. The paper also outlines evaluation metrics, cross-dataset protocols, and publicly available datasets, highlighting the scarcity of realistic in-vehicle benchmarks. Finally, we discuss deployment challenges unique to vehicular settings and identify future directions such as domain-adaptive learning, synthetic and simulation-based data generation, and multimodal large language models. By bridging current methods with the specific demands of vehicles, this survey provides a foundation for developing secure, lightweight, and explainable FAS mechanisms suitable for deployment on intelligent vehicle systems.","PeriodicalId":73307,"journal":{"name":"IEEE transactions on biometrics, behavior, and identity science","volume":"8 4","pages":"497-517"},"PeriodicalIF":4.4,"publicationDate":"2026-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148433290","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

SCOTT-Gait: A Spiking Cross-Attention Temporal Triple-Loss Network for Occlusion-Compensated Gait Recognition scott -步态：一个脉冲交叉注意时间三重损失网络用于闭塞补偿步态识别

IF 4.4

IEEE transactions on biometrics, behavior, and identity science Pub Date : 2026-07-01 Epub Date: 2026-04-03 DOI: 10.1109/TBIOM.2026.3680495

Ying Tao;Shengyu Gao;Yue Zheng;Chip-Hong Chang

{"title":"SCOTT-Gait: A Spiking Cross-Attention Temporal Triple-Loss Network for Occlusion-Compensated Gait Recognition","authors":"Ying Tao;Shengyu Gao;Yue Zheng;Chip-Hong Chang","doi":"10.1109/TBIOM.2026.3680495","DOIUrl":"https://doi.org/10.1109/TBIOM.2026.3680495","url":null,"abstract":"Occlusions significantly degrade gait recognition performance, particularly when only anonymized representations like binary silhouettes are provided instead of original videos due to privacy concerns. Existing methods either rely on computationally expensive silhouette reconstruction or extract occlusion-invariant features, which often underperform under severe occlusions. To address these limitations, we propose SCOTT-Gait–a Spiking Neural Network (SNN) that integrates spatial and temporal occlusion compensation mechanisms directly into the recognition model for accurate gait recognition under occlusions. Spatial compensation is achieved through a deformable spiking-based convolution, which infers missing structural information by modeling the co-movement between visible and occluded regions within a binary silhouette image. These adaptively sampled, geometrically correlated samples are processed by the SNN backbone augmented with a Temporal Cross-Attention (TCA) module, which is designed to address temporal compensation by enhancing the network’s ability to aggregate temporal features. A Cycle-Consistency Triplet Loss (CCTL) is introduced to further align features across the walking period to ensure that similar poses are grouped closer together in the feature space, even if they are from different points of the walking cycle. Extensive experiments on both synthetic and real-world occluded gait datasets demonstrate that our method maintains competitive performance compared to state-of-the-art CNN-based approaches, while remaining computationally efficient - a remarkable milestone as the first vision-based SNN gait recognition with occlusion. In particular, it has the most robust performance when trained on naturally, mildly occluded samples and evaluated on severely occluded synthetic samples.","PeriodicalId":73307,"journal":{"name":"IEEE transactions on biometrics, behavior, and identity science","volume":"8 4","pages":"652-665"},"PeriodicalIF":4.4,"publicationDate":"2026-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148433311","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0