Gauss-Hermite moments application in medical image segmentation

Medical image segmentation is a challenging problem. In this paper, a new automatic approach for MR brain image segmentation is presented. It is based on properties of Gauss-Hermite moments (GHMs) and fuzzy c-means (FCM).First, GHMs filter and GHMs detection are introduced and applied to image in preprocessing stage. FCM then performs to segment white matter (WM) and gray matter (GM). Subsequently, subtraction on GHMs detection is utilized to extract cerebrospinal fluid (CSF). Finally, examples on T1-weighted MR brain image are presented and show the efficiency of this approach.