MULTIMODAL BIOMETRIC AUTHENTICATION WITH FACE, FINGERPRINT AND IRIS PATTERNS USING ENHANCED SEGMENTATION AND FEATURE EXTRACTION

Multi-biometric authentication integrates more than one biometric trait, sample, sensor, or algorithm in the identification and verification process. Due to its advantage of increased reliability, robustness, speeding up of the identification process, it’s rapidly attaining attractive emerging trend in the global biometrics market. The proposed multimodal biometric approach for human identity recognition with the enhanced method of feature extraction and classification process uses the integration of face, finger print and iris texture patterns. With the aim of increasing the processing speed and achieving the level of confidence of recognition outcomes of the multimodal biometric system, Fuzzy logic feature selection employing Fuzzy Bacterial Foraging algorithm is employed for the features extracted using Convoluted Local Tetra Pattern. Adaptive median filtering provides contrast enhancement to view image noise free. In the classification juncture, Relevance Vector Machine (RVM) classifier is used to obtain superior classification performance than SVM, Neural network and HMM classifiers. Iris segmentation is achieved by means of Weight Sampled Geodesic Active Contour technique, which diverge from customary Geodesic Active Contour scheme. The simulation outcome evaluation on MATLAB with FVC2004 – Finger Print, CASIA – Iris and MIT-CBCL - face image datasets outperforms the existing methods akin to Gabor filter, Neural Network, HMM, 2D Wavelet, DCT based approach bearing in mind the parameters of recognition time taken, true positive, true negative, accuracy, sensitivity and specificity.