基于眼底图像视网膜血管分割、分类方法的中医目诊研究

Research on TCM Eye Diagnosis Based on the Segmentation and Classification of Retinal Vessels in Fundus Images

中医目诊是通过观察眼部的神、色、形、态来诊断全身疾病的一种方法。对眼底血管的观察是目诊的主要内容。视网膜血管结构的变化与许多慢性疾病密切相关。图像处理技术在眼底图像研究的应用可以为系统性疾病提供早期诊断。眼底图像血管分割是眼底血管研究的基础,目前大体可分为基于匹配滤波、血管跟踪、形态学处理、形变模型以及机器学习的分割算法5类。眼底血管分类,如动静脉分类,为临床诊断提供重要价值。在此类研究的基础上,提出了眼底血管分割、分类方法在中医目诊中的实际应用方法。

Eye diagnosis of traditional Chinese medicine(TCM)is a method to diagnose systemic diseases by observing the changes of spirits,expressions,shapes and states of eyes.Fundus observation is the most important content of eye diagnosis.The changes of retinal vascular structure are closely related to many chronic diseases.The application of image processing techniques to fundus images can provide early diagnosis of systemic diseases.Fundus image segmentation is the basis of fundus vascular research and can be roughly divided into five main categories:matched filtering,blood vessel tracking,mathematical morphology,deformable model,and machine learning.Classification of fundus vessels,such as arteriovenous classification,provides important value for clinical diagnosis.Based on this kind of research,we put forward the practical application methods of fundus segmentation and classification methods in the eye diagnosis of TCM.