An Eyelashes Segmentation Method Based on Improved Inter-Class Variance Maximization Algorithm

Abstract

Iris extraction is a crucial step in iris recognition technology. However, it is easily interfered with by eyelashes. In the process of iris recognition, the detection of eyelashes is very significant when taking iris images. But the precision of the existing iris detection algorithm is not high. This paper proposes an eyelash segmentation method based on the adaptive threshold. Firstly, a specific area in the picture is selected as the region of interest according to the position of the eyelashes. Then the gray range of the area is defined according to the iris gray information. Finally, under the above two constraints, the optimal threshold of gray-scale image segmentation is calculated using the algorithm of maximizing variance between classes. The method improves the subjective accuracy of eyelash segmentation in the iris image and lays the foundation for the next step of iris recognition.