Classification ofMR Brain Tissues Using Fuzzy Estimation

A detailed analysis procedure is described for segmenting T1-weighted volumetric MR brain into different tissue types based on fuzzy classification. The main aim in this study is to compensate for the blurring effect on tissue boundaries due to partial volume effects. The method used in this paper is described as follows. First, the intracranial volume (ICV) is separated from the scalp and skull with a deformable contour model. The model requires a set of control points (CPs) to start with which roughly delineate the volume of interest on several slices. The starting CPs are regenerated by using the cubic spline interpolation to produce the same number of evenly spaced CPs and then sorted so that they are roughly adjacent to each other in the third dimensions. Splines are then generated between each set of connecting CPs in the third dimension. Finally, new CPs are then constructed by normalizing the energy values to move the CPs to the desired regions. Second, in order to deal with the problem of the partial volume effect, an algorithm for fuzzy segmentation is presented which has integrated fuzzy spatial affinity with statistical distributions of image intensities for each of the three tissues - WM, GM and CSF. This algorithm is tested on well-established simulated MR brain volumes to generate an extensive quantitative comparison with different noise levels and different slice thicknesses ranging from 1 mm to 5 mm. Finally, the results of this algorithm on clinical MR brain images are demonstrated.