基于外推跟踪的眼部白睛血管自动提取方法  

Automatic Vessel Extraction for Sclera-conjunctiva Images Based on Exploratory Tracking

作　　者：朱贵冬[1] 沈理[1] 王今觉[2] 

机构地区：[1]中国科学院计算所 [2]中国中医研究院,北京100700

出　　处：《计算机工程》2005年第17期6-8,共3页Computer Engineering

基　　金：2000年中国中医研究院院所长基金资助项目(中研科2000(378))

摘　　要：提出了一种从中医“望目辨证”眼部图像提取血管的方法。首先利用最优阈值分割和形态滤波,提取白睛区域,且动态设置等距扫描线并在其上寻找血管跟踪初始点。跟踪时通过探测当前中心点两侧边缘点走向,结合血管剖面分析,动态地调整跟踪向量的方向和长度。血管提取可自动进行,并可处理曲率和直径呈非渐进变化的血管形态。This paper presents algorithms for vessel extraction in sclera-conjunctiva images, which can be applied in syndrome differentiation by observing eyes （traditional Chinese medicine）. Sclera-conjunctiva region is isolated by optimal thresholding segmentation and mathematical operation as a first step. The scan lines in sclera-conjunctiva region are searched for the initial tracking points. Tracking vector is updated according to the current vessel edge and profile analysis. The results indicate that this fully automatic method could overcome the problems of discontinuity in vessel curvature and diameter.

关 键 词：医学图像处理 眼部自睛血管提取 自适应外推跟踪 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]