Abstract

According to Iridology, the shape and position of iris plaques can reflect the health condition of human organs. In this paper, we propose a method to detect iris plaques automatically with computer vision technology. Firstly, we find the alternative regions of the plaques by searching the iris image with a window. In these regions, the mean of the central pixels’ gray level is less than the minimum of the marginal pixels’ gray level. Secondly, with Level set method, we obtain the edge information of these alternative regions. Finally, we determine whether the contour of the zero level sets meets the conditions of closed criterion and dimension criterion or not. If met, treat the pixels in the contour as the edge of the plaques. By comparing with the other algorithms, we found that our method is more efficient and suitable to complex background as iris images and it’s less blind.

• Xiao Nan Liu
• Wei Qi Yuan

Reference: https://link.springer.com/chapter/10.1007%2F978-3-642-25449-9_13