Biometric authentication has gained prominence as a secure and user-friendly alternative to traditional authentication systems. However, unimodal biometric systems often face limitations such as spoofing, poor data quality, and environmental variations. This paper presents an AI-powered multimodal biometric fusion system that integrates facial recognition and fingerprint matching verification. Leveraging deep learning techniques such as FaceNet, Siamese EfficientNet, and ResNet-HOG architecture, our framework performs feature-level and score-level fusion to enhance security, robustness, and accuracy. Evaluated on benchmark datasets,our system achieves a recognition accuracy of 99.81%for face, 90%for fingerprint, 89.1% for signature. The proposed system shows high potential in secure authentication applications such as e-governance, financial services, and access control.
Madhiraju et al. (Thu,) studied this question.