Multi-Modal Biometric Recognition System: Fusion of Face and Iris Features using Local Gabor Patterns

Abstract

Biometrics, described as the science of recognizing an individual based on her physiological or behavioral traits, is beginning to gain acceptance as a legitimate method for determining an individual’s identity. Multimodal biometric system utilizes two or more individual modalities, e.g., face, gait, iris and fingerprint, to improve the recognition accuracy of conventional unimodal methods. Multimodal biometric systems overcome problems such as noisy sensor data, non-universality or lack of distinctiveness of the biometric trait, unacceptable error rates, and spoof attacks by consolidating the evidence obtained from different sources. In this paper, we have developed an efficient technique for multimodal biometric recognition using the face and iris images. In our proposed technique, features from face and the iris images are extracted and the features from both the modalities are concatenated to form a combined feature vector, which also contains the number of irrelevant pixels in the iris image. The extraction process is done utilizing both the local Gabor patterns and the LBP to form LGXP (Local Gabor XOR Patterns). For recognition, the combined feature vector of a face and iris image are extracted and is compared with the database. The average matching score is calculated, which is based on the distance measure and also on the given weightage based on the irrelevant pixels. Based on the average matching value, the decision is to be made whether the test image is recognized or not. For experimental evaluation, we have used the face and iris image databases and the results clearly demonstrated that the proposed technique provided better accuracy in biometric recognition. Keywords:- Biometrics, Multi-modal biometrics, Face Recognition, iris recognition, Gabor feature, LBP operator (Local Binary Pattern), Local Gabor XOR Patterns.