Automatic segmentation of tumor, EDEMA and healthy tissues of brain using neuro fuzzy inference system

In this study we segmented brain MR images into constituent tissues using a neuro-fuzzy inference system. We segmented T1, T2 and FLAIR MR images of the patients with glial tumor into white matter, gray matter, cerebrospinal fluid and diseased tissues as edema and tumor. We retained only brain tissue in the images by stripping the skull that is beyond the scope of this study. We used statistical features obtained from the stationary wavelet transform coefficients as the input of the system. We compared the results obtained from the system with the manually segmented tissue areas and evaluated using Dice similarity index. We have shown that neuro fuzzy inference system shows quite high performance for segmenting brain MR images and can be used effectively for this purpose.