Feature Level Fusion of Iris and Sclera using Entropy Based CNN Features to Improve the Performance of Biometric Authentication
International Journal of Advanced Trends in Computer Science and Engineering · 2020
Today biometric system are commonly used for person authentication based on physical and behavioral biometric modalities like iris, face, finger prints, ear, sclera, DNA, voice, signature, etc.Instead of using standalone biometric system, multimodal biometric systems are secure and provide more accurate results for person identification and verification.This paper describes the multimodal eye biometric system where iris and sclera features are extracted using CNN based on entropy values to perform the accurate automatic segmentation.Feature level fusion is performed using color and texture characteristics of iris and pupil with Y-shaped sclera characteristics from eye image based on support value.Unconstrained color eye image database UBIRIS.v2 and MMU are used for experimentation and testing on MATLAB platform.The proposed eye biometric system outperform in case of segmentation and recognition accuracy.Segmentation accuracy 97.8% for iris, 98.1% for sclera and 99.4% for pupil is achieved for UBIRIS.v2database.Recognition accuracy is 97.99% for unconstrained eye image UBIRIS.v2 and 93.33% for NIR image database MMU.