基于Level set方法的医学图像分割

本文引入了一种基于偏微分方程的曲线进化方法—Level set方法,通过与Fast marching方法的结合,可以实现运算速度的大大提高。同时引进了更有效的Kim提出的GMM(Group Marching Method)方法,减少了运算量,并给出了改进方法。最后,把该方法用于仿真图与医学图像分割中,获得了较好的效果。

Level Set Method Based Medical Image Segmentation

A PDE (Partial Differential Equation)-based curve evolution method, i.e. Level-set method, is proposed. Using this method combined with fast marching, the operational speed can be greatly improved. A new method called GMM (Group Marching Method), which was presented by Kim, was preferred because of its efficiency. Furthermore, an improved method was also given in this paper. Examples of experiment using this method for both the synthetic image and the medical image are illustrated. At last, encouraging results were shown in medical image segmentation. It seems that the method proposed has better performance than those presented before.