SEGMENTATION OF MULTIPLE-SCLEROSIS LESIONS IN INTENSITY CORRECTED MULTISPECTRAL MRI

To segment brain tissues in magnetic resonance images of the brain, we have implemented a stochastic relaxation method which utilizes partia l volume analysis for every brain voxel, and operates on fully three-d imensional (3-D) data. However, there are still problems with automati cally or semi-automatically segmenting thick magnetic resonance (MR) s lices, particularly when trying to segment the small lesions present i n MR images of multiple sclerosis patients. To improve lesion segmenta tion we have extended our method of stochastic relaxation by both pre- and post-processing the MR images. The preprocessing step involves im age enhancement using homomorphic filtering to correct for nonhomogene ities in the coil and magnet. Because approximately 95% of all multipl e sclerosis lesions occur in the white matter of the brain, the postpr ocessing step involves application of morphological processing and thr esholding techniques to the intermediate segmentation in order to deve lop a mask image containing only white matter and Multiple Sclerosis ( MS) lesion. This white/lesion masked image is then segmented by again applying our stochastic relaxation technique. The process has been app lied to multispectral MRI scans of multiple sclerosis patients and the results compare favorably to manual segmentations of the same scans o btained independently by radiology health professionals.