Brain tumor segmentation using multiphase level set based on bias field correction

This paper presents a method for abnormal brain tissues segmentation in presence of intensity inhomogeneity. A variety of region-based algorithms have been proposed for this problem. Most of them require an edge term to add local information. The proposed approach is based on a region level set model which is performed on a bias corrected field. In addition, it involves three steps: first, a noise removal and skull removal method is used for pre-processing the magnetic resonance image; then, a bias corrected fuzzy c-means algorithm is utilized for getting a bias corrected field; finally, multiphase level set method is applied on corrected domain to segment the brain tumor. Results demonstrate that the proposed method is very effective for the segmentation of tumor in clinical magnetic resonance image.