Thresholded and morphological operations based brain image segmentation

Abstract

Fig. 1 Flowchart of TMOS The masking operations with reference to the chosen RoI, facilitates the selection of pixels confined within the boundary of RoI. The pixels bounded by the RoI are preserved while the remaining pixels are discarded using the ellipseConvolution filter,. This convolution operation enables to separate the RoI from MRI image, leaving out the remaining pixels (i.e) unwanted pixels. In general the morphological image operators viz.,erosion, dilation, opening and closing are applied on digital images separately or is combination, in order to refine the process of segmentation using structuring element of varying size, as required. The TMOS uses opening and closing for this purpose ofskull stripping on the brain image obtained out of RoI selection. The opening of A by B is computed as the erosion of A by B, which is suceeded by dilation. The opening of A by B is: AB = (AB)B (1) Morphological closing is similar to dilation in which the boundaries of foreground regions are based on the given structuring element. The closing of A by B is: A•B = (AB)B (2)