基于颜色相似系数的虚拟人股肌分割

虚拟可视化人研究是虚拟人研究计划的初始阶段,而虚拟人可视化人研究的必备前提是快速、准确地进行彩色图像数据集的组织分割.引入一种足球机器人系统中快速目标定位的算法,在RGB颜色空间中,利用两个像素之间的颜色相似系数的计算结果,同时根据设定的阈值来判断两种颜色是否具有相似性.该方法省去了在其他针对彩色图像的分割方法中,颜色空间的转换过程,节约了大量计算时间,提高了图像分割过程的效率.实验结果证明,该方法可对彩色图像进行有效的分割,符合人类视觉特性,特别适合于虚拟人计划中海量彩色尸体切片的组织分割.

Segmentation of thigh's muscle for virtual human based on color similarity coefficient

The study of visualized virtual human is the beginning stage of the virtual human project.And the necessary precondition of visualized virtual human research is the success of fast and exact segmentation that aims at the color image data set.In this paper,a segmentation method that is often used for the fast recognizing target in soccer robot system is imported.In the color space of RGB,the result of color similarity between two pixels is used to judge that whether they have the comparability on the base of the threshold enacted.This method leaves out the conversion process of color space in other ways of segmentation for color image.It saves a lot of time of calculation and increases the efficiency of segmentation.The experiment result proves that this method can segment the color image effectively,and accords with the human's vision identity.It is especially propitious to the organic segmentation for the plentiful color image of virtual human project.