A Novel Iris Segmentation Approach Based on Superpixel Method

Iris recognition system is under a rapid development of biometric systems because of the uniqueness and high reliability of iris. The segmentation of iris is a key step in iris recognition, which has a significant effect on the quality of subsequent feature extraction and matching. The main propose of iris segmentation is to get more iris region and exclude eyelashes, eyelids and other interference effectively, and the process speed must meet the requirement of a real time recognition system. To address those problems, a novel segmentation approach is proposed based on super pixel method. Firstly, SLIC algorithm is used in image segmentation, then we extract normalized histograms as super pixel features. Finally, the correlation distance is applied to measure the similarity between two adjacent super pixels. This process is iterative to converge to get the final segmentation. Experimental result shows that the proposed method is effective on removing interference, and get a more complete iris image.