Iris detection and marking based on Haar features

This paper presents an iris detection and marking method based on Haar features from images of figures. We firstly apply Haar-features detection algorithm and Canny edge detection algorithm, to the image, then scan and extend the contours in particular areas to get an initial U-Curves marking iris contours, and then carries on sample expanding and noise reduction to get optimized sample points. At last, the iris contour with optimized sample points is fitted to an ellipse using least squares method, thus get the iris position. Experimental results show that our approach U-Curves improve the accuracy of iris detection and marking.