Brain tumor CT image segmentation based on SLIC0 superpixels
Xiaopeng Wang, Peng Ma, JunJun Zhao · 2016
The brain tumor CT image segmentation frequently suffers from the fuzzy edges, and manual segmentation mainly relies on doctor's clinical experience. For the purpose to accurately segment brain tumor, a method for brain tumor CT image segmentation based on SLIC0 superpixels is proposed. Firstly, the simple linear iterative clustering version with 0 (SLIC0) is employed to generate superpixels; Secondly, region merging is adopted to merge the similar superpixels according to their gray, and finally segment the brain tumor regions. Experiments show that this method can accurately segment the target tumor, and segmentation accuracy can be adjusted by setting the pixel number of superpixels.