融合FCA和积分图的CT图像分割-《计算机技术与发展》

文章信息/Info

Author(s):: LI Hong-feng; LIU Hong-shen; BU Ying-min; YE Yin-qiu; Computer College, Anhui University of Technology

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

Abstract:: 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