基于颜色空间变换的数字化人体切片图像背景移除算法  

Background Removing Algorithm of Cross-sectional Images of Digitized Visible Human Dataset Based on Color Space Conversion

作　　者：沈颖[1] 宋文强[1] 

机构地区：[1]第三军医大学基础医学部计算机教研室,重庆400038

出　　处：《华南国防医学杂志》2006年第2期18-21,29,共5页Military Medical Journal of South China

基　　金：国家自然科学基金资助项目 (NO:30270698)

摘　　要：目的对数字化人体图片进行预处理,完整提取其中的人体数据。方法应用首例中国数字化可视人体数据集,提出一种基于颜色空间变换并引入标记矩阵进行移除背景的算法,并以头部切片为例进行详细阐述。结果将原始数据从RGB颜色空间变换到YIQ颜色空间后,对I灰度图取阈值为f=0进行标记,再结合标记矩阵对原始图像进行背景移除的操作,能够完好地提取出所需要的人体数据。结论本文所提出的算法简单易行而且运算效率高,对目标区域的数据保存较好,为图片的下一步处理奠定了基础。Objective To pre - dispose the visible human picture, and to extract the integrated human data from the hole images. Methods The color space of the Cross -sectional images of fresh tissues from the Chinese visible human dataset is converted（ the head data was taken as an example） , and their background was removed with a mark -matrix. Results After converting the color space from RGB into YIQ, the I gradation image was marked by the threshold being zero. The background was well removed from the primary images by the mark - matrix. Conclusions The algorithm proposed in this paper is high feasible and efficient. The integrate data extracted are useful for the next procedure of the visualization of the digitized visible human dataset.

关 键 词：背景移除 颜色空间 Ⅰ分量 标记矩阵S 

分 类 号：TN911.73[电子电信—通信与信息系统]