Skull-Stripping MR Brain Images Using Anisotropic Diffusion Filtering and Morphological Processing

The traditional methods for MRI image skull-stripping are not precise. Aiming at this problem, this paper develops method based on anisotropic diffusion filtering and morphological processing. Firstly, it uses anisotropic diffusion filter to smoothes noisy regions in the image. Secondly, the edge detector is used to locate the anatomical boundaries. Lastly, the task of the morphological processor is to select the pixel corresponding to the brain tissue from the original image. The results show that this method can segment brain precisely. The experiments show that this method can segment the brain and non-brain precisely, realize skull-stripping.