Iris Image Recognition based on Combined Hamming and Cosine Distances Approach

Abstract

A combined method comprising Hamming and Cosine distances for iris recognition is proposed in this article. In the proposed method, image pre-processing is accomplished by binarization method. Feature extraction is performed based on localization and thresholding principles. Finally, recognition is performed through Hamming and Cosine distances matching principles between pixels. Iris dataset from Chines University of Hong Kong (CUHK) are considered for experimental analysis. Experimental results prove that the Hamming distance method provides 94% of matching accuracy between different person’s iris images and Cosine distance method provides 100% matching accuracy with same person’s iris images. Overall, this hybrid iris recognition system provides 94% efficiency.