Segmentation and extraction of GBM tumor in brain MRI medical images : comparative study

Abstract

In this study, a comparison between two image segmentation methods has been discussed; the first method is based on normal brain's tissue recognition then tumor extraction using Thresholding method. The second method is classification based on fuzzy EM segmentation, which is used for both brain recognition and tumor extraction. Medical image examinations often use more information that are acquired from multiple imaging modalities, this use is increasing due to the complementary information that can be obtained from data fusion. The later improves the quality of the diagnosis; for instance, the image fusion can be developed for data from different modalities or different individuals, it may also concern fusion of image data with an external model, this can express prior knowledge about the problem at hand. The image data used in this comparative study belong to Algerian patients with Glioblastoma multiform. Since the goal of these methods is to detect, segment, extract, classify and measure properties of the brain normal and abnormal (tumor) tissues, the results of comparative study provided a guide tool of which method is more accurate in term of GBM volume estimation for the studied samples of the patients