A multimodal performance evaluation on two different models based on face, fingerprint and iris templates

In this paper, we present a multimodal face recognition system that fuses results from both Principal Component Analysis, Fisherface projections, minutia extraction and LBP feature extraction on different biometric traits. The proposed identification system uses the face, fingerprint and iris of a person for recognizing a person. We use two different methods for comparing the performance. The first model used principal component analysis to extract the features of the fingerprint and iris image and fisherfaces for the face image. The second method used fisherface for face, minutiae extraction for fingerprint and LBP feature for iris image. The developed multimodal biometric system possesses a number of unique qualities, starting from utilizing principal component analysis and Fisher's linear discriminant methods for individual matchers identity authentication and utilizes the novel feature fusion method to consolidate the results obtained from different biometric matchers. Two fusion strategies are experimentally compared. The proposed approaches is tested on a real database consisting of 500 images and shows promising results compared to other techniques. The Receiver Operating Characteristics also shows that the proposed methods are superior compared to other techniques under study.