Integrative Fusion Paradigms in Multimodal Biometric Authentication: A High-Precision Framework Leveraging Multi-Trait Synergy for Robust Human Identification

Abstract

This study presents a robust multimodal biometric recognition system integrating face, ear, iris, and foot traits. Using PCA, Eigen images, Hamming distance, and Haar transforms, trait-specific features were extracted and fused at score, rank, and decision levels. The system was validated on a 100-person self-created dataset, achieving recognition accuracy up to 96%, significantly outperforming unimodal approaches. Score-level fusion with logistic regression reduced the EER to 3.2%, enhancing decision reliability. Practical applications span national ID systems, border control, and secure device authentication. Fusion of complementary modalities addressed issues of spoofing and intra-class variability. The study demonstrates high adaptability across environments and data types. Advanced techniques like PSO and CNNs further boost precision and scalability. This research highlights the growing feasibility of secure, efficient, and user-friendly biometric systems for real-world deployment.