A robust comparative study of five brain extraction algorithms (BET; BSE; McStrip; SPM2; TMBE)

The segmentation of the brain is one of the preprocessing steps performed neuroimaging laboratories, and many brain extraction algorithms (BEA) have been developed to meet this need and to make it automatic to cope with the huge number of MRI must be treated every day. In this paper we propose a comparative study of famous different algorithms of brain extraction most popular and recent methods proposed, namely BSE (Brain Surface Extractor), BET (Brain Extraction Tool), SPM2 (Statistical Parametric Mapping v2), McStrip (Minneapolis Consensus Strip) and the new method TMBE (Threshold Morphologic Brain Extraction). This will be validated by comparison of synthetic and real MRI using different criteria for evaluation. This comparison may then play the role of a guide for future researchers.