说明：  无边活动轮廓模型(C-V模型)是水平集分割方法中的一种经典模型.传统的无边活动轮廓模型将灰度同质作为区域分割准则,这使其对于仅含两个同质区域且灰度变化不大的图像能够取得很好的分割效果,但对灰度渐变图像分割时,该模型往往无法得到正确结果.本文针对这一问题,通过引入Chebyshev距离构造一种新的相似度,以此来表征演化曲线内外灰度差异,修改了传统无边活动轮廓模型中均值取值的定义,使得演化曲线在图像灰度变化缓慢区域获得较大的驱动力.新模型克服了传统无边活动轮廓模型不能正确分割灰度渐变图像的不足.实验对比及分析表明,新模型能够更准确地分割灰度渐变图像,同时对噪声有一定的鲁棒性.
( Boundless active contour model (CV model) is the level set segmentation method in a classical model of traditional active contour model endless gray homogeneous region segmentation as a criterion for making it with only two homogeneous regions and gray little change in image segmentation can achieve good results, but gray gradient image segmentation, the model often can not get the correct results.)