摘要： 针对多相C-V模型计算复杂性高、对初始轮廓敏感等问题，提出一种结合区域梯度的活动轮廓模型。该模型对演化曲线内部区域使用梯度进行拟合，对演化曲线外部区域使用灰度进行拟合。实验结果表明，相对于多相C-V模型，该模型能快速分割多相图像，对初始轮廓不敏感。

关键词: 图像分割, 活动轮廓模型, 水平集方法, 偏微分方程, C-V模型, 多相C-V模型

Abstract: Chan-Vese(C-V) model is computationally expensive and sensitive to initialization. In order to address these problems, a new active contour model combined with region gradient is proposed, which uses the gradient information in the interior of contour and the intensity information in the exterior of contour. Experimental results show that the proposed model can more quickly segment multi-phase images and allows more flexible initialization compared with the multi-phase CV model.

Key words: image segmentation, active contour model, level set method, partial differential equation, Chan-Vese(C-V) model, multi-phase C-V model

中图分类号: 

• TP911.73