AI-Driven Secure Authentication: A Deep Learning Approach for Multi-Modal Biometric Systems
A. Jagan, Sharmila Bee. S, K. Monisha, V. Velmurugan, S. Priyadharshni, B. Jegajothi · 2025
Biometric authentication technologies have become quite popular owing to their great accuracy and security. However, specific biometric features such as face, hand, and iris pictures often have limits when utilized alone. To address these problems, in this study offer the MMFusion-Net model, which combines three biometric features (facial, hand, and iris) using a Deep Learning-based fusion technique. The model blends cuttingedge Convolutional Neural Networks (CNNs) with innovative fusion processes to increase biometric identification systems' accuracy and dependability. MMFusion-Net outperforms standard models such as SVM, MLP, CNN, and others in terms of accuracy, precision, recall, and F1-score. The model's efficacy was further verified using comprehensive hyperparameter tweaking and ablation tests, which confirmed the importance of multimodal fusion and deep learning architectures. The findings emphasize the need of integrating different biometric features to create strong and secure identification systems. MMFusion-Net outperforms previous approaches and establishes itself as a potential alternative for future biometric applications.