Iris region extraction based on a new combined method
The human iris is a complex textured entity. Iridology, also known as iris diagnosis, is a subject which is the study of determining the body organ’s health through the inspection of iris texture. The iris is an internal organ of the eye, which is located behind the cornea and in front of the lens. Publications related to segmenting the iris region constitute a significant fraction of the published work in iris biometrics. Iris segmentation (iris localization or iris edge detection), which aims to isolate the actual area of the eye image, is an important stage of iris recognition. Iris segmentation algorithms that assume circular boundaries for the iris region continue to appear in some conferences. Some iris segmentation methods which were focused on removing the assumption of circular boundaries were proposed. [Z. He,2009, S. Shah,2009] Some researchers [R.D. Labati,2009; E.P. Wibowo,2009; J. Zuo, 2010; Ning Wang, 2012), 8334] have considered various approaches to segmenting the iris with boundaries not constrained to be circles. A number of researchers prefer to use methods based on Hough-transform to segment the iris edge from image, they are Ma [L. Ma, 2004], Huang [J. Huang, 2004], Lim [S. Lim, 2001], and Yuan [X. Yuan, 2005] respectively. Iris image contains abundant types of textures, most of them are visible with naked eyes and their structural features are quite obvious. In this section, we present how to efficiently represent the iris textures and measure the features of different kinds of textures. About the texture methods, Wildes [R. Holonec, 2006] proposed an iris texture analysis method, Boles [A. Discant, 2006] proposed an iris recognition method based on the detection of
zero-cross points. These are three most recognized iris recognition methods and the Daugman’s is the most applied one [J. Daugman, 2002]. Tamura texture was selected to present the plaque characteristic. In the most cases, only the first three Tamura’s features are used for the CBIR. These features capture the high-level perceptual attributes of a texture well and are useful for image browsing. However, they are not very effective for finer texture discrimination [V. Castelli, 2002]. For this reason, we proposed a combined method. The object region was pre-segmented by tamura texture method, and the watershed transform was used to segment the boundary of iris. The combination of watershed segmentation and texture feature can resolve the weaknesses.