Fingerprint and Face-Based Secure Biometric Authentication System Using Optimized Robust Features

Abstract

Security holds an integral position in every field. Numerous security measures and recognition system have been implemented to enhance the security aspects. In this pape, the authors have proposed two biometric recognition systems to deal with the security and authentication systems. Fingerprint recognition system (FPRS) is based on minutiae feature extraction of fingerprint image and in face recognition system (FRS). Viola–Jones algorithm (VJA) is implemented for face detection from static images. Image features are extracted using speeded up robust features (SURF) that is further optimized using genetic algorithm (GA). Recognition efficiency of the proposed systems is improved by training and classification of the optimized features based on feed-forward back-propagation neural network (FFBPNN). These unimodal biometric recognition systems are evaluated in terms of confusion matrix parameters, precision, TDR, f-measure and detection accuracy. Simulation results have established that the systems demonstrated an average recognition accuracy of 91.25% (FPRS) and 92.28% (FRS). Moreover, it has also been established that by increasing a sample size from 10 to 500 images, the recognition accuracy of the FPRS gets enhanced by 28.8% and FRS by 19%. Additionally, FRS outperformed FPRS by exhibiting 0.84% higher detection accuracy.