基于双向轮廓跟踪的面部轮廓线生成算法  

Generation algorithm of facial contour lines based on bidirectional contour tracking

作　　者：郑丽萍[1] 李光耀[2] 薛凌燕[1] 

机构地区：[1]聊城大学计算机学院,山东聊城252059 [2]同济大学CAD研究中心,上海201804

出　　处：《计算机工程与设计》2010年第2期447-450,454,共5页Computer Engineering and Design

基　　金：国家自然科学基金项目(60771065)

摘　　要：为了提高医学图像中面部轮廓线的分割效果,提出一种新的面部轮廓线生成算法(bidirectional contour tracking,BCT)。该算法首先利用边缘检测算法对CT图像进行检测,然后利用头部CT图像中各组织的灰度特征和图像的对称性,对检测后的图像进行双向轮廓跟踪,并获得面部的轮廓线。另外,根据面部轮廓线的特点,使用中点法可生成辅助轮廓线。以一组颅脑CT图像为例来验证算法的效果,实验结果表明,新的生成算法提高了轮廓线的生成速度,并有很好的分割效果。To improve the segmentation effect of facial contour lines in medical image, a novel generation algorithm of facial contour lines, which is called bidirectional contour tracking （BCT）, is proposed. At first, edge detection algorithm with specific threshold is used to detect CT image. Then, according to the features of the grayscale of head CT image and the image symmetry, bidirectional contour tracking algorithm is used to segment images which is detected. At last, facial contour lines are generated. Otherwise, according to the feature of facial contour lines, midpoint method is used to generate auxiliary contour lines. In experiment, take a group of head CT images as example, and verify this algorithm efficiency. Experimental results show that this novel algorithm can elevate the generating speed of contour lines and create good segmentation effect.

关 键 词：图像分割 灰度值 边缘检测 轮廓跟踪 辅助轮廓线 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]