Hough Transform and Chi-Square-Based Iris Recognition

The paper proposes a new iris recognition method based on canny edge detection, Hough transform, pseudo-polar coordinate, wavelet transform and chi-square algorithms. The iris image is firstly converted to greyscale before the edge detection, segmentation, normalization, feature extraction and feature matching operations are performed in sequential order. Feature extraction module uses wavelet transform-based approach to extract standard iris features while the feature matching module relies on chi square algorithm for reference and template-based feature matching. Analyses of the experimental results revealed satisfactory performance of the new method and its dependence on the quality of the hardware used for image acquisition. Comparison of obtained results with those from some existing iris recognition systems showed competitiveness and superiority of the new method