边缘强化的无需重新初始化的水平集分割模型

Level set segmentation model with enhanced boundary and non-reinitialization

对Chan-Vese模型和Li等提出的不需初始化的基于变分的几何活动轮廓模型在水平集框架下的物理机制进行了分析,在考虑两种模型优缺点的基础上,提出一种新的基于水平集框架的图像分割模型.该模型整合了图像边缘的局部信息和区域的全局信息,数值计算过程中水平集不需要重新初始化.为了防止边缘信息深入到分割目标的内部,新模型利用Laplacian修正算子加大边缘信息在方程中的权重.实验表明,与CV模型相比,所提出的新模型分割效果和分割时间与初始轮廓线的位置和形状选取基本无关;在处理噪声图像、灰度值渐进多目标图像和边缘复杂图像等效果也优于CV模型和Li模型.

In this paper CV model and the geometric active contour model based on variation provided by Li are analyzed. A new level set model based on PDE is proposed to overcome the drawbacks of these models above - mentioned. The new model combines the local information of edges and the global information of regions, and there is no necessary of re - initialization in numerical computing. Experimental results show that, compared with CV model, the segmentation effect and the running time of the new model are independent of the position and shape of the initial contour. The effect of the improved model was much better than the two models above -mentioned when processing noisy images, multi -object images and images with complex edges.