Paper Title
HYBRID FUSION STRATEGIES FOR MULTIMODAL BIOMETRICS: INTEGRATING FACE, IRIS AND FINGERPRINT USING YOLOV7Abstract
The rapid growth of digital services and security-sensitive applications has increased the demand for reliable, accurate, and robust biometric identification systems. Single-modal biometric systems often suffer from limitations caused by variations in illumination, pose, image quality, occlusion, noise, and spoofing attacks. To overcome these challenges, this study proposes a hybrid fusion framework for multimodal biometric identification that integrates facial, iris, and fingerprint characteristics using advanced deep learning techniques and the YOLOv7 object detection architecture. The proposed approach exploits the complementary information contained in multiple biometric modalities to improve identification accuracy and system reliability. In the proposed framework, biometric samples are first acquired and preprocessed to reduce noise and enhance discriminative features. YOLOv7 is employed for efficient and rapid detection and localization of relevant biometric regions, particularly facial and ocular regions, while specialized feature extraction mechanisms are applied to fingerprint and iris patterns. The proposed methodology emphasizes improved accuracy, robustness, computational efficiency, and generalization capability compared with conventional unimodal and basic multimodal biometric systems. The integration of face, iris, and fingerprint traits provides a richer representation of individual identity, while YOLOv7 contributes to fast and precise biometric-region detection.
KEYWORDS : Multimodal Biometrics, Face Recognition, Iris Recognition, Fingerprint Recognition, Hybrid Fusion, Feature-Level Fusion, Decision-Level Fusion, YOLOv7, Deep Learning, Biometric Identification.
