Iris recognition using texture analysis and band limited phase only correlation
Vibha Vyas, M.S. Sewani · 2009
Iris recognition is an emergent and important biometric identification technique. This paper describes two different methods for iris recognition, namely texture analysis, and band limited phase only correlation. The body of this paper details steps for iris segmentation, normalization, and matching. Iris segmentation and normalization are common for both the methods. In texture analysis, the normalized image is filtered using spatial filter whose kernel is suitable for iris recognition and feature vector is generated from the filtered coefficients. While, BLPOC uses phase components in 2D DFT of normalized image to match two iris images. FAR and FRR graphs are plotted for both the methods. Comparison of the present algorithms is done with existing methods. Such a comparison will be useful for future work. (4 pages)