Iris Edge Extraction and Location by Advanced Watershed and Active Contour Method

Abstract

The final performance of iris recognition system is partly dependent on the accuracy of iris segmentation (iris localization or iris edge detection). A new iris location method based on watershed transform and an active contour model was proposed in this paper. To avoid over-segmentation problem, watershed transform with a marking function and minima imposition was proposed. The active contour model converged to the precise edge from the initial contour obtained by the watershed transform. The method we used can improve the speed of the image segmentation. It was applied on 200 iris images with 90.04% of the test targets successfully segmented. And the processing time of an image is less than the others. Algorithm proposed in this paper proved the combination method is an effective method for segmentation of iris boundary.