（苏州大学 智能信息处理及应用研究所, 江苏 苏州 215006）
关键词：边界跟踪; 方向记忆; 八邻域; 脑血管图像
Rapid boundary tracking algorithm based on cerebrovascular image feature
SUN Xiao-ping, WU Jian, ZHAI Hai-tao, CUI Zhi-ming
（ Institute of Intelligent Information Processing & Application， Soochow University， Suzhou Jiangsu 215006, China）
Abstract:When dealing with complex brain vascular images, classic boundary tracking algorithm exposed some disadvantages such as boundary was not high enough smooth, accuracy was limited and its speed was unsatisfied. This paper proposed a new rapid boundary tracking algorithm to solve those problems. Analysed the feature of cerebral vascular image edge that vertical details were more than horizontal details. Based on this feature, determined next border point with memory direction of the search options and different weights to different search directions. Experimental results show the proposed algorithm can get smooth boundary and its performance is enhanced greatly. For the next phase of cerebral vascular shape and feature extraction, the new algorithm provides accurate data.......