TY - CHAP T1 - Multiple Sclerosis lesion quantification in MR images by using vectorial scale-based relative fuzzy connectedness T2 - Medical Imaging 2004: Image Processing Y1 - 2004 A1 - Ying Zhuge A1 - Jayaram K Udupa A1 - László Gábor Nyúl ED - J Michael Fitzpatrick ED - Milan Sonka AB - This paper presents a methodology for segmenting PD- andT2-weighted brain magnetic resonance (MR) images of multiplesclerosis (MS) patients into white matter (WM), gray matter (GM),cerebrospinal fluid (CSF), and MS lesions. For a given vectorialimage (with PD- and T2-weighted components) to be segmented, weperform first intensity inhomogeneity correction andstandardization prior to segmentation. Absolute fuzzyconnectedness and certain morphological operations are utilized togenerate the brain intracranial mask. The optimum thresholdingmethod is applied to the product image (the image in which voxelvalues represent T2 value x PD value) to automaticallyrecognize potential MS lesion sites. Then, the recently developedtechnique -- vectorial scale-based relative fuzzy connectedness --is utilized to segment all voxels within the brain intracranialmask into WM, GM, CSF, and MS lesion regions. The number ofsegmented lesions and the volume of each lesion are finally outputas well as the volume of other tissue regions. The method has beentested on 10 clinical brain MRI data sets of MS patients. Anaccuracy of better than 96% has been achieved. The preliminaryresults indicate that its performance is better than that of thek-nearest neighbors (kNN) method. JF - Medical Imaging 2004: Image Processing PB - SPIE CY - Bellingham; Washington N1 - ScopusID: 5644264947doi: 10.1117/12.535655 ER -