multimodal biometric fusion

Existing biometric systems face a fundamental trade-off between multi-modal recognition capability and user convenience, as single frontal-face images typically lack sufficient information for robust multi-feature extraction. To address this, we propose a lightweight multi-modal biometric framework that simultaneously extracts five distinct biometric traits—face, iris, periocular region, nose, and eyebrows—from a single input frontal-face image. This work presents the first end-to-end joint modeling approach for five modalities from one image, implemented via a multi-branch deep neural network with feature-level fusion. We validate the method on the CASIA-Iris-Distance dataset. Compared to conventional single-modal and multi-source multi-modal baselines, our approach achieves a +3.2% improvement in identification accuracy while preserving acquisition simplicity and significantly enhancing cross-scenario robustness. The proposed framework establishes a new paradigm for low-intrusion, high-security biometric authentication.

To address the poor generalization and weak robustness of single 3D face reconstruction (3DFR) models in unconstrained surveillance scenarios, this paper proposes the first systematic multi-algorithm collaboration framework. Our method synergistically integrates heterogeneous 3DFR approaches—including CNN-based methods, deformable 3D Morphable Models (3DMM), and neural radiance fields (NeRF)—to generate complementary 3D representations. We design a dual-path score-level fusion strategy: one parametric (weighted/Bayesian) and the other non-parametric (ranking/decision-tree-based). Additionally, we establish a unified evaluation framework covering cross-distance, cross-device, and cross-dataset settings. Extensive experiments under diverse real-world conditions demonstrate that our approach achieves an average 12.7% improvement in identification accuracy and reduces cross-dataset generalization error by 31%, significantly enhancing the reliability and adaptability of 3D face verification.

Existing biometric suitability assessment frameworks—such as the 1998 comparative table—are severely outdated, failing to reflect technological advances and emerging security threats. Method: We propose the first modern, multimodal suitability reassessment framework integrating expert judgment with empirical evidence. It combines structured surveys of 24 domain experts with uncertainty-aware modeling and cross-modal consistency analysis across 55 publicly available biometric datasets. Contribution/Results: Our framework quantifies dynamic shifts in accuracy, security, and usability across major modalities—e.g., improved face recognition performance but declining fingerprint reliability. Expert consensus is high and strongly aligned with empirical findings. Crucially, we identify previously unrecognized methodological gaps and key points of scholarly disagreement. This work establishes a verifiable, updatable assessment paradigm for biometric selection and pinpoints critical research directions, including robustness enhancement and adversarial-resilient design.

To address the challenges of large template sizes and high computational overhead of homomorphic encryption in privacy-preserving multimodal biometric recognition, this paper proposes a training-free, interpretable feature truncation-based dimensionality reduction method. The method performs modality-adaptive truncation directly on deep neural network–extracted features (face, fingerprint, and iris) prior to encryption, significantly compressing template size while preserving discriminative information. Its design is natively compatible with homomorphic encryption, enabling secure cross-modal fusion and matching. Experiments on a self-constructed virtual multimodal database demonstrate that the fused templates achieve a 67% size reduction, while the equal error rate (EER) outperforms all single-modal baselines—without compromising recognition accuracy or security; in fact, both are maintained or improved. This work is the first to introduce training-free feature truncation into privacy-preserving multimodal biometrics, achieving an effective balance among computational efficiency, recognition accuracy, and practical deployability.

To address low facial resolution, identity (ID) confusion, and background distortion in multi-identity full-body portrait generation, this paper proposes FPGA—a plug-and-play end-to-end framework. Methodologically, it introduces (1) DDIM Inversion-driven ID Repair Inference (DIIR), the first approach to decouple precise facial detail restoration from background fidelity; (2) a Multimodal Fusion (MMF) training strategy that enhances target-ID representation specifically within facial regions; and (3) IDZoom, a million-scale multimodal dataset, coupled with RepControlNet for accelerated inference. Experiments demonstrate that FPGA achieves state-of-the-art performance across both objective and subjective metrics in multi-ID scenarios. On a single L20 GPU, inference completes in ≤2.5 seconds, enabling high-fidelity face swapping and cross-style ID transfer. The framework exhibits strong generalization and practical applicability.

This study addresses the lack of compliant and practical threshold calibration methods for face verification systems in border control under extremely low false match rates (FMRs), hindered by legal and privacy constraints on real-world data acquisition. It presents the first systematic evaluation of synthetic facial data for calibrating identity-to-live face verification thresholds in high-security scenarios. Through score distribution alignment, cross-domain threshold transfer, and adversarial morphing attack testing, the work demonstrates that synthetic data can approximate real-data calibration behavior under controlled conditions. However, under uncontrolled settings, performance degrades significantly due to tail distribution mismatches, introducing notable security vulnerabilities. The findings reveal that calibration efficacy is highly dataset-dependent, highlighting the limited generalizability of synthetic data in real-world deployment contexts.

This study addresses the dual challenge in face verification systems: defending against presentation attacks (e.g., photos, videos) while maintaining robustness to legitimate appearance variations such as accessories, lighting, and pose changes. The authors propose a unified framework that, for the first time, integrates classical handcrafted features—including PCA, LBP, HOG, SURF, and Harris corner detectors—within five fusion strategies: Product Matching (PM), Linear Product Matching (LPM), Hierarchical Product Matching (HPM), Sum Matching (SM), and Hybrid Matching (HM). These strategies jointly optimize spoof detection and recognition robustness across preprocessing and classification stages. Experimental results demonstrate that HPM achieves 94.59% accuracy under mixed spoofing attacks, 81.5%–93.2% under lighting and pose variations, and 91.67% anti-spoofing performance; LPM yields the best anti-spoofing rate (93.2%) but exhibits weaker pose robustness. The work further reveals a quantifiable sensitivity–robustness trade-off between spoof detection and appearance invariance.

This study addresses the challenge that complex backgrounds in unconstrained scenarios—such as airport border control—degrade face recognition accuracy and impair the detection of presentation attacks. The authors systematically evaluate the impact of multiple face segmentation methods on four representative recognition models and three attack detection techniques through comprehensive experiments on datasets encompassing both controlled and unconstrained imagery. For the first time, they comprehensively demonstrate the dual role of background removal in simultaneously influencing recognition performance and security mechanisms, showing its significant effects on image quality, identification accuracy, and attack detectability. These findings provide empirical grounding and practical guidance for preprocessing strategies in real-world biometric systems, effectively bridging the critical gap between deployment feasibility and system reliability.

This work addresses the poor compatibility between contactless 3D fingerprint systems and conventional 2D fingerprint recognition, as well as the challenge of cross-modal matching. To overcome these limitations, the authors propose a non-parametric unified framework that does not rely on a global finger model. By fusing multi-view 3D point clouds and employing pose-aware unwrapping, elliptical fitting for pose normalization, and a cross-modal registration strategy, the method achieves high-fidelity conversion and precise alignment from 3D fingerprints to their 2D equivalents. Evaluated on a self-collected multimodal database of 150 fingers, the system demonstrates a 3D fusion error of approximately 0.09 mm and ridge-level 3D–2D registration accuracy, significantly enhancing matching performance in real-world scenarios.

This study addresses the fluctuation of model discriminability across identities in open-set person re-identification by proposing a training-free, identity-conditioned score fusion framework. Specifically, this work introduces a parameter-free fusion rule that uniquely integrates both query and gallery identity conditions. By exploiting intra-identity consistency and cross-identity impostor comparisons to extract specific profiles, the framework adaptively customizes fusion weights for each gallery identity. Without requiring additional training, the proposed method significantly enhances the separability between genuine and impostor matches. Extensive experiments demonstrate that it consistently outperforms existing baselines across three clothing-change benchmarks, achieving an absolute reduction of up to 8.8% in the false non-identification rate.