基于双水平集的图像分割模型

针对水平集模型对于具有细长拓扑部分的目标和弱边界目标进行分割时存在的问题，提出了双水平集方法.在新的方法中通过两条水平集之间的相互吸引来加速解的收敛，同时提出了一种快速有符号距离函数生成方法，提高了计算效率.传统的水平集通常利用图像边界信息来构造速度函数进行求解，但在待分割目标具有很强噪音或具有弱边界时往往得不到真实解，对此，提出了一种新的基于区域信息的速度构造方法.将双水平集模型应用到合成图像与左心室MR图像的分割实验，结果表明该方法具有较好的分割效果和较高的分割效率.

An Image Segmentation Model Based on Dual Level Sets

It is hard for level set method to segment an image with slight topological structure. By analysing the phenomenon, a new method based on dual level sets has been introduced. With this new method, objects can be segmented quickly. Meanwhile, typical level set method can not get the true segmentation when it segments an image with strong noise or week boundary, for it only uses edge information to construct the speed function. So the region information is employed to construct the speed function. Better results are achieved in the application of this method on segmentation of synthetic or cardiac magnetic resonance images.