Multi-Modal Biometric Authentication System Using Hybrid Convolutional Neural Networks (HCNN) Based on Face, Finger Vein and Iris Fusion

The increasing demand for accurate and secure biometric authentication has spurred the creation of complex multimodal biometric systems. A Hybrid Convolutional Neural Network (HCNN) is suggested in this study as a multi-modal biometric system that integrates face, vein, and iris recognition. The technique uses deep learning techniques to enhance feature extraction and fusion, increasing identification accuracy and resistance to spoofing attacks. The proposed system employs federated learning to ensure privacy - preserving model training and encryption for data transfer and storage, therefore addressing significant privacy and security concerns. Additionally, liveness detection techniques are offered to ensure the accuracy of the biometric data and prevent spoofing efforts. To ensure data-security, AES-265 is used for encryption. Experimental results show considerable improvements in recognition accuracy of 96.5% and precision of 97.1%, reduced error rates, and a significant reduction in processing time, indicating that the system is suitable for large-scale deployment. With applications in a number of industries, such as financial services and access control, this technique offersa complete, reliable solution for safe and effective biometric authentication.