Brain Tumor Type Detection Using Texture Features in MR Images

In this paper, the algorithms for the detection of brain tumor and then classifcation of the tumor into meningioma and glioma are proposed. Firstly, automated method is proposed for skull stripping using mathematical morphology and thresholding. Stationary wavelet transform features, Self-organizing map (SOM) and watershed algorithm are used for the segmentation of brain tumor. Gray level co- occurrence matrix (GLCM) features are extracted from tumor and feed forward neural network is used for classification. Proposed algorithm reported classification accuracy of 95% with the available dataset of real brain images from the hospital.