A Data-Adaptive Fuzzy Rule base System for Putamen Segmentation in Brain MR Images

Abstract

Abstract TRecent researches reveal that the abnormality in human putamen area is closely related to many neurological diseases. Magnetic resonance (MR) imaging is the most common methodology in such studies. In order to measure the volume of putamen in MR image, accurate segmentation method is highly desirable. In practice, the inhomogeneity effect in MR imaging process generates gradual changes in white-to-gray matter contrast in putamen area, which makes segmentation task rather difficult. In the present study, an adaptive fuzzy rule base (FRB) system is proposed and applied to both simulated and real MR images for putamen segmentation. Compazing to FSL (FMRIB Software Library)’s Automatic Segmentation Toolbox (FAST), K-means and Expectation Maximization (EM) algorithm, the proposed adaptive FRB system demonstrates better performance in putamen segmentation.