基于蚁群算法的甲状腺结节超声图像边沿检测法  

Edge Extraction of Thyroid Nodule in Ultrasound Images Based on Improved Ant Colony Algorithm

作　　者：朱玲[1] 施心陵[1] 刘亚杰[1] 田溪[1] 

机构地区：[1]云南大学信息学院电子工程系,昆明650092

出　　处：《计算机工程》2006年第24期178-179,239,共3页Computer Engineering

基　　金：云南省自然科学基金资助项目(2002C0002Z)

摘　　要：甲状腺结节超声图像一般较为模糊不易判断识别,该文基于蚁群算法,提出了一种新的甲状腺结节超声图像边沿检测方法。针对超声图像特点,算法对信息素及食物源的设置进行了改进,阐述了该算法,同时将检测结果与Canny算法和Sobel算法的结果进行了比较,实验证明该改进的蚂蚁算法提取出了清晰的甲状腺结节的边沿。The thyroid nodule in the ultrasound images is usually blur and difficult to discern and analyze. A new algorithm which is based on the ant colony algorithm （ACA） for extracting the edges of thyroid nodule in ultrasound images is presented. Taking the characters of ultrasound images into account, the method of setting up the pheromone and the food-source is developed and the algorithm is elaborated, Compared with other image segmentation algorithms, such as the Canny operator and the Sobel operator, ACA can achieve better segmentation results.

关 键 词：甲状腺结节 蚁群算法 信息素 边沿检测法 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]