Iris feature extraction method based on LBP and chunked encoding

Abstract

Iris feature extraction is a key issue in iris recognition. This paper proposes a novel iris feature extraction method based on local binary pattern (LBP) images and the chunked encoding method. Firstly it applies the LBP to the normalized iris image and obtains the iris' LBP image, then extracts the iris's feature via the chunked encoding method based on the iris' statistical information. Finally it completes the iris recognition and classification using Hamming distance. Experimental results showed that this algorithm can get higher recognition rate than the traditional iris feature extraction method, which demonstrated the efficiency of the proposed method.