A Frequency Based Approach on Biometric Identification System Using Multiple Traits of Face and Iris

In the modern era, biometric identification place an important role to identify humans in a unique manner. Single or multiple physical traits can be used in various applications. Single Biometric traits face problems such as poor environment, nonuniversality, noisy data. To overcome these, multimodal biometric identification uses more than one physical trait so that it increases the performance in identification which is not possible in single biometric system. Also, it is difficult for the intruder to simultaneously spoof the multiple traits. This paper uses face and iris as multiple physical traits. To extract the features of face and iris Local Binary Pattern is used and phase only correlation for latter. The fusion of features is done using frequency based fusion where Log Gabor filter is used in matching score level.