A New Image Segmentation Method Combined with Anisotropic Diffusion
Luo Cheng-jun · Jisuanji fangzhen · 2010
Active Contours without Edges(CV model) has been widely used in image segmentation,especially in detecting objects whose boundaries are not defined by gradient.However,the model is sensitive to noise and fail to deal with range images.In this paper,we cast anisotropic diffusion into CV model to remove noise and preserve important features such as edges.CV model with the term of without re-initialization is transferred into a new image segmentation method.Various experimental results implemented via Matlab show that the model is more robust to noise and can detect images with intensity inhomogeneity or complex background.Also,it can effectively segment range images while the CV model is not applicable.