Performance analysis of partitioning-based and subpattern-based approaches on iris recognition

Abstract

This paper presents the performance analysis of partitioning-based and subpattern-based methods on iris recognition without applying the traditional iris detection methods. We propose a simple and efficient partitioning-based approach for iris recognition using non-overlapped partitions on the iris images and applying feature extraction methods on these partitions to recognize the irises. These partitions are individually experimented and then the output of each partition is combined using a multiple classifier combination method. In this respect, original PCA and subspace LDA methods are used as feature extractors with the combination of the preprocessing techniques of histogram equalization and mean-and-variance normalization in order to nullify the effect of illumination changes which are known to significantly degrade recognition performance. The recognition performance of the partitioning-based approaches is compared with the performance of subpattern-based PCA and subpattern-based subspace LDA approaches in order to demonstrate the performance differences and similarities between these two types of approaches. To be consistent with the research of others, our work has been tested on three iris databases namely CASIA, UPOL and UBIRIS. The experiments are performed on these three iris databases to demonstrate the recognition performances of the proposed partitioning-based approaches, subpattern-based approaches and traditional PCA and subspace LDA approaches.