Research on iris segmentation based on level set method with keeping circular
Lina Fan · Jisuanji yingyong yanjiu · 2014
Iris segmentation is one of the most important part of iris recognition system,the good or bad of its segmentation will affect the accuracy in iris recognition,and which is one of the most reliable human biological identification for life.This paper proposed an algorithm about iris segmentation based on level sets.The combination between level set implicit characteristics and shape constraint with explicit equation ensured that the evolving curve kept circular in the process of evolution to segmentation inner edge.In addition,it introduced the adaptive area term to CV model with shape constraint to stop outer edge.The experimental results show that,although the limit of opening the eyes,glasses and eyelashes block and imaging equipment problems form the image point of view,the proposed model can still obtain a good segmentation effect.It selects the regional overlap degree DICE as segmentation evaluation index,and the experimental results show that the proposed algorithm is effective for iris segmentation.