In Figure 18, the interpolated image has been obtained from the pre-processed brain MRI after
resizing. Normalized cut algorithm with marked segment boundaries technique has been applied
in the interpolated image. From the segmented image it can be visualized that the WM, GM and
CSF are distinguishable with minimum quality. However, the segmentation can be improved
with further investigation of this technique. Only two segmented groups were obtained like
white matter region and gray matter-CSF region.
An original image has been segmented with varying segment numbers in Figure 19. When the
segmentation number of the normalized cuts algorithm is low, the brain tissues- GM, WM and
CSF can be differentiated to a little extent. But if the segmentation number is increased beyond
100 levels the quality of the segmentation becomes worse to identify. Basically, the area of the
image is divided rather segmenting the brain tissues in this method.