Eyelid and eyelash detection method based on morphology

The reliable detection of the occlusion iris image in the polar domain is required to improve the performance of iris recognition. Estimated the occluding and refine a much more accurate iris mask before iris feature extraction is essential step in the iris recognition. In this paper, we propose a method for detecting occlusion and refining mask accurately based on morphology. Compared with traditional method which detected sequence is eyelid-eyelash and distinguish eyelash based on threshold alone. Out method segment image by using the connectivity of the eyelash which is the base of the region grow. It guards against the noise in the process of the thresholding and decreases the computer load significantly. This method is applied to the CASAI version 3 databases and shows good performance.