Iris recognition based on grouping KNN and Rectangle Conversion

In iris recognition, as a large amount of experiments show, the inner edge of iris is not a standard circle, thus edges may cause the error of accurate recognition. If we use traditional localization method of round template, it can cause the problem of iris legacy, the loss of iris textures and longer time as well. To improve the accuracy of iris location, reduce the recognition time, this paper develops a new iris recognition algorithm. Firstly, the lights pot within the pupil is filled in the original image, then the image is unfolded into a rectangle and the circle detection is substituted by the point and line detection in the rectangle image to find the inner and outer edge, secondly, texture features are extracted by EMD. Thirdly, the K nearest neighbors (KNN) of each test sample are found based on distance of Mahalanibis. Lastly, recognition results are decided by majority voting method. The recognition accuracy of simulation experiments based on CASIA iris image database amounts to 99% and has the less running time. The results show that compared to circle template, Rectangle Conversion has more accurate location of the iris, thus effectively raising the recognition accuracy.