An Iris Recognition Algorithm for Identity Authentication

Iris recognition is very accurate and stable,it has been widely applied in finance, border control and access control,etc. The paper proposes the iris recognition algorithm for identity authentication. The algorithm includes iris image preprocessing, iris boundaries-location and iris segmentation, normalization, iris texture feature extraction and matching. Firstly, the iris image is pretreated by median filter and improved salt and pepper algorithm to reduce noise. Second, the iris is modeled by gray projection method for rough position, and the inner boundary of the iris is precisely positioned by point Hough transform(PHT), while the outer boundary of the iris is precisely positioned by the improved Daugman algorithm. Third, the algorithm based on the line segment extraction is used to normalize the image after the position and the wavelet transform is used to extract the feature after normalization. Finally, the two iris are identified by calculating the Hamming distance of the two iris codes. The experimental results show that the accuracy of the system is 97.8%.