一种面向中医目诊的虹膜定位算法

摘要

The iris localization has an important supplement function for eyes diagnosis of traditional Chinese medicine. The correlation of iris and its image characters with the factors such as disease, constitution, elements of traditional Chi-nese medicine syndrome differentiation, should be thoroughly studied. In this paper, an iris localization algorithm for TCM eyes diagnosis is proposed. The approximate position of the pupil center is obtained by gray level threshold estimation, and then the Hough transform is improved via parameter tuple merging to localize the inner boundary of the iris. Then, the K-means clustering is used to estimate the approximate range of the iris region, and the initial contour is constructed. Finally, The active contour model is used to obtain the outer boundary of the iris, which avoids the failure of Hough transform in lo-calizing iris with serious image distortion.%虹膜诊断对于中医与民族医学的目诊理论具有重要的补充和借鉴作用,有必要深入研究虹膜及其图像特征与疾病、体质、中医辨证的证候要素等因素间的对应关系.本文以基于中医目诊的虹膜特征分析为出发点,以虹膜仪采集图像为对象,研究并提出了一种基于中医目诊的虹膜定位算法.该算法首先采用基于灰度阈值估计的方法得到瞳孔中心大致位置,再基于相似参数组的合并改进霍夫变换算法,得到虹膜内边界的定位结果,解决虹膜仪图像的瞳孔收缩、边界发生部分形变的问题.然后采用K均值聚类估计虹膜区域的大致范围,并以此构造初始轮廓,利用主动轮廓模型得到虹膜的外边界,从而避免由于图像畸变而引起霍夫变换失效的问题.