Canny算子联合阈值分割的解剖图像背景去除  

Background Removal Based on Canny and Thresholds Complex Segmentation Technology

作　　者：宁旭[1] 张绍祥[2] 谭立文[2] 李晓寒[1] 

机构地区：[1]第三军医大学物理教研室,重庆400038 [2]第三军医大学解剖教研室重庆市计算医学研究所,重庆400038

出　　处：《中国医学物理学杂志》2007年第5期326-328,355,共4页Chinese Journal of Medical Physics

基　　金：国家自然科学基金资助(No.60473128)

摘　　要：研究数字化人体图像的背景去除,是数字化可视人体图像的研究领域(chinese visible human)中必不可少的一步。本研究提出了一种基于Canny算子计算目标边缘联合阈值分割有效地去除了数字化人体图像的背景。算法首先利用归一化阈值分割技术排除大部分背景,再结合边缘跟踪对目标边界附近保留的边缘进行识别、连接、去噪,准确、有效的去除了背景。用Matlab软件对上述方法进行验证,通过与其他几种分割算法的比较,本方法实际使用效果较好。To study on the background removing in the digital image of human body. After removing most of the background with standardized thresholds segmentation, the authors do identification, connection and noise-removing work against the edge remained around the target verge, and then the results of these two methods are combined to get the final image. Complex Segmentation Technology Based on Edge and Thresholds effectively removes the background in the image of human body. Compared with other segmentation algorithm, the algorithm presented here has better practical effect.

关 键 词：背景去除 图像分割 薄层断面解剖 

分 类 号：R319[医药卫生—基础医学]