融合FCA和积分图的CT图像分割  

A CT Image Segmentation Algorithm by Using FCA and Integral Image

作　　者：李宏峰[1] 刘宏申[1] 卜应敏[1] 叶荫球[1] 

机构地区：[1]安徽工业大学计算机学院,安徽马鞍山243032

出　　处：《计算机技术与发展》2012年第6期127-129,142,共4页Computer Technology and Development

基　　金：国家自然科学基金项目(61003311)

摘　　要：鉴于目前FCA的演化规则有大量计算,迭代一次的耗时比较长,而积分图具有优化运算速度的特点,文中将积分图和FCA相结合,提出融合FCA和积分图的CT图像分割方法。该方法以FCA为基础,用多值图像来完成元胞状态的初始化,设计了一个完成边界提取和加权去噪的模糊演化规则,对CT图像进行迭代演化,最终得到分割图像。整个过程中,利用积分图算法来进行优化,达到单次迭代耗时减少的效果。实验结果表明该方法性能上可以达到图像分割的效果,而且在耗时方面比普通演化规则的耗时要低。Based on FCA＇ s evolution rules have large number calculation and long time consuming of each iteration, while integral image has the characteristics of optimizing computing speed, a new segmentation algorithm which integrates FCA with integral image is proposed in this paper. The method is based on FCA,using multi-value image to initialization of cellular,designing fuzzy evolution rules that can complete edge extraction and weighted denoising, and then utilize fuzzy evolution rules to iteration of CT image gray-scale and get the final segmentation image. Throughout process, use the integral image algorithm to optimize and reduce time-consuming of each iteration. Experimental results show that the method has good advantage of time-consuming and performance.

关 键 词：FCA 积分图 CT图像 图像分割 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]