Aug 2026· F1000Research· Vol 15, pp. 1367· 0 citations· 34 references
TL;DR
These findings suggest that pairing transfer learning with GAN-based augmentation is a practical, effective way to overcome data scarcity in finger vein recognition, offering a promising direction for dependable biometric systems.
Abstract
Background Finger vein biometrics has evolved into a reliable biometric modality because of its inherent resistance to spoofing and robustness to external attacks. In recent years, advances in deep learning-based person identification using finger vein trait have further enhanced its performance and effectiveness. However, convolutional neural networks (CNNs) need large volumes of labeled data to perform well, and finger vein datasets are typically small, making this shortage of training samples one of the major obstacles to building reliable, deployable systems. Method This paper addresses that gap by combining transfer learning with Conditional GAN-based data augmentation to build a more robust finger vein recognition framework. Rather than relying on conventional augmentation alone, synthetic, class-conditioned finger vein images are generated using two GAN variants, Conditional DCGAN and Conditional WGAN-GP, which incorporate class-conditional information to enable stable, controllable adversarial training and improve intra-class compactness and inter-class discrimination while enriching the training data. To evaluate the generalization capability of the proposed approach, extensive experiments are conducted on two widely used benchmark databases, THU-FVFDT2 and FV-USM, under various data-split schemes. For classification, four established CNN architectures such as VGG19, MobileNetV2, InceptionV3, and EfficientNetV2-M were fine-tuned and evaluated using standard classification metrics for all partitioning schemes. Results The experimental results were consistently strong across both datasets and confirming that the framework generalizes well with consistently high classification accuracy across both datasets. As observed, MobileNetV2 reached 99.74% identification accuracy on THU-FVFDT2 dataset, while EfficientNetV2-M achieved the best overall accuracy 99.85% on FV-USM dataset using the Conditional DCGAN augmentation technique. Conclusion These findings suggest that pairing transfer learning with GAN-based augmentation is a practical, effective way to overcome data scarcity in finger vein recognition, offering a promising direction for dependable biometric systems.
OpenVeinNet, a finger vein verification framework designed for cross-dataset and open-set evaluation, is presented and it is shown that explicitly modelling local vein geometry, global vascular relationships, and angularly compact embeddings is effective for openset finger vein verification.
Sushrut Patwardhan, Raghavendra Ramachandra· IEEE Transactions on Biometr...· 0 citations
Optimize Deep Learning–based Adversarial Defense Mechanism (ODL-ADM) is proposed in this work, which projects adversarial samples into an immune feature space that is both discriminative and resistant to perturbations.
Sheilla Ann Bangoy Pacheco, Mahesh Goyani, Jayzel P. Bangoy et al.· ITEGAM- Journal of Engineeri...· 0 citations
The reliability of finger vein biometric systems is increasingly threatened by sophisticated presentation attacks. Current presentation attack detection (PAD) methods, often relying on supervised learning, are vulnerable to novel, unseen attacks because they depend on comprehensive labeled spoof datasets that are impractical to collect. To address this zero-day threat, this research proposes a one-class anomaly detection framework trained solely on authentic finger vein samples. While utilizing a standard U-Net-inspired denoising autoencoder as the architectural backbone, this work introduces two key novel contributions to tailor the model for biometric security: (1) a Multi-Objective Anomaly Detection Loss Framework that uniquely integrates multi-scale reconstruction error, gradient preservation constraints, and deep Support Vector Data Description loss to strictly regularize the latent space and (2) a Spectral Error Optimization technique that applies adaptive frequency weighting to amplify subtle texture artifacts inherent in spoof mediums. This combination is significant because the multi-objective loss forces the model to learn fine-grained physiological vein patterns, while the spectral optimization captures high-frequency anomalies often missed by spatial reconstruction alone. Experimental results on the FVPAD-USM dataset demonstrate that this approach achieves an Attack Presentation Classification Error Rate below 2% and an Average Classification Error Rate below 10%. By outperforming supervised baselines like support vector machines and convolutional neural networks in generalizing to unseen digital and glossy photo attacks, this work establishes the effectiveness of unsupervised, frequency-enhanced anomaly detection for robust biometric security.
Received: 30 August 2025 | Revised: 18 March 2026 | Accepted: 18 June 2026
Conflicts of Interest
The authors declare that they have no conflicts of interest to this work.
Data Availability Statement
The data that support the findings of this study are openly available in the FVPAD-USM database at http://drfendi.com/fvpad_usm_database/, reference number [33].
Author Contribution Statement
Mohd Shahrimie Mohd Asaari: Conceptualization, Methodology, Software, Formal analysis, Writing – original draft. Bakhtiar Affendi Rosdi: Validation, Investigation, Writing review & editing, Supervision. Andrew Tiong Hoe Pin: Methodology, Software, Validation, Investigation, Data curation, Visualization. Zahid Ur Rahman: Investigation, Resources, Data curation. Muhammad Firdaus Akbar: Validation, Funding acquisition.
M. Asaari, B. Rosdi, Andrew Tiong Hoe Pin et al.· Artificial Intelligence and...· 0 citations
Comparing the deep learning architectures used for biometric verification in a fair manner has remained a difficult task. Researchers typically rely on different datasets, employ varying preprocessing steps, and adopt inconsistent evaluation criteria, all of which muddy the picture when trying to decide whether one network truly outperforms another. This work attempts to cut through that noise by fixing every variable except the backbone architecture itself. Residual Network-50) ResNet50), Efficient Network Version 2 Small (EfficientNetV2-S), and the Swin Transformer-Tiny (Swin-T) were trained and tested on the same LUTBIO and XJTU datasets, applying identical subject-disjoint splits across face, fingerprint, and palmprint modalities. The analysis measures Equal Error Rate (EER), Area Under the Curve, and True Acceptance Rate at False Acceptance Rate = 0.1%, and further examines five distinct score-level fusion strategies across every possible two- and three-modality pairing. The single-modality experiments reveal a clear pattern: Swin-T achieves a perfect 0.000% EER on faces, ResNet50 handles fingerprints best at 1.957% EER, and EfficientNetV2-S leads on palmprints with 0.140% EER. On LUTBIO, 10 out of 12 tested fusion configurations reach a reported EER of 0.000%. This study reports the effect sizes and selected confidence intervals for the key comparisons, allowing both their statistical and practical significance to be assessed. The results show that no single architecture is universally superior; the best-performing backbone is modality-dependent. Score-level fusion consistently improved verification accuracy over the best single modality, substantially reducing error rates. These findings imply that architecture choices for biometric systems should be modality-aware and that score-level fusion is an appropriate method for robust multimodal verification. The controlled protocol additionally provides a reproducible framework for fair architectural comparison in future biometric research.
Rowida Jasim Alazawi, N. Habeeb, Alaa Jabbar Qasim Almaliki· Kurdistan Journal of Applied...· 0 citations
This review presents a comprehensive analysis of recent deep learning and transfer learning techniques for fake image detection, examining widely adopted convolutional neural network architectures, benchmark datasets, evaluation metrics, and current research developments.
Nisha Parveen, Anjali Saxena· International Journal for Re...· 0 citations
The widespread use of biometric identification systems has raised important issues regarding template security and scalability. In contrast to passwords, biometric characteristics, once exposed, cannot be withdrawn, thus raising permanent risks of identity exposure. Moreover, traditional systems require complete retraining when adding new users, thus causing computational inefficiency and low scalability. This work presents an original framework for cancelable multi-biometric identification that combines face and iris dynamics in an incremental deep learning model. The proposed method produces non-invertible and revocable templates through deep feature extraction by ResNet-50, followed by a random projection transformation, thus meeting the ISO/IEC 24745:2022 security requirements. A dynamic 1D-CNN model, enhanced by elastic weight consolidation and rehearsal learning, enables incremental learning without catastrophic forgetting or complete model retraining. Experimental evaluation shows an average recognition rate of 98.98% on incremental datasets, with false acceptance rates as low as 0.115% and true acceptance rates of 98.93%. The proposed framework provides an optimal trade-off among security, scalability, and recognition performance, thus filling the most important gaps in current biometric systems and offering a basis for the development of a next-generation privacy-preserving identification infrastructure.
Shaik Badulla, Vundavalli Balasankar, G. L. V. Prasad· 2026 4th International Confe...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
