双重主动轮廓图像分割  

Double Active Contour Evolution for Image Segmentation

作　　者：王卫卫[1] 梁伟文[1] 水鹏朗[2] 

机构地区：[1]西安电子科技大学理学院,西安710071 [2]西安电子科技大学雷达信号处理重点实验室,西安710071

出　　处：《工程数学学报》2007年第3期458-462,共5页Chinese Journal of Engineering Mathematics

基　　金：国家自然科学基金(60272058).

摘　　要：本文首先利用小波变换得到原图像的粗分辨逼近,在粗分辨逼近中得到图像的一个粗尺度分割。由于逼近图像中噪声下降,尺寸减少,使得算法对参数的选取不太敏感,而且收敛速度加快。然后将第一次分割结果通过小波反变换返回到原始尺度上,将得到的近似轮廓曲线作为初始水平集函数再在原图像中演化得到更准确的分割,这样就得到了一种双重主动轮廓图像分割算法。由于初始轮廓曲线非常接近真实的轮廓曲线,所以很快就可以收敛到真实的轮廓。理论分析和数值结果表明,双重主动轮廓分割算法可以快速有效地分离出感兴趣目标。In this paper, the image is first transformed into the wavelet domain to get a coarse approximation, on which the first active contour evolution is performed. Since the approximation image contains less noise and its size is a fraction of that of the original one, the active contour evolution converges fast and the result does not depend heavily on the involved parameter. Then the resulted image is inverted to the spatial domain, from which we get an approximation contour. The approximation contour is taken as an initial level set function and the second active contour evolution is performed on the original image to get the real contour, In this way, a double active contour evolution algorithm for image segmentation is derived. As the approximation contour is very close to the real contour, the second evolution converges very fast, Experiment results show that our method is fast and efficient in image segmenting.

关 键 词：图像分割 主动轮廓 水平集 小波变换 

分 类 号：O175.2[理学—数学] TP391[理学—基础数学]