Development Of Non Circular Iris Segmentation Using Modified Snake Algorithm

Abstract

Listen

Translate article

The importance of this study is to identify a human’s almost round iris shape since the existing technique of iris recognition detects human iris using a circular segmentation method. However, the circular segmentation has limitations to detect and acquire iris in non-ideal conditions. In the non-ideal condition, the location of the iris is situated at a different angle and the information acquired may not be about the iris template. Thus, the non-ideal condition produces less accurate information about the iris template and creates a “not match” situation during verification since the information of the iris template acquired is not similar to the previous iris template. To overcome the problem, this paper proposes the development of the Modified Snake algorithm. The experiment results show that the Modified Snake algorithm reduced the influence of noise in iris images, good execution in learning the edge of images since the number of iteration reduced or maintain based on the iris image since iris image contains high noise. As a conclusion, the experiment used the public iris image database from the Center for Biometrics and Security Research for non-bias evaluation and demonstrates an efficient performance.