Brain MRI Segmentation and Bias Estimation via an Improved Non-Local Fuzzy Method

Intensity in homogeneities cause considerable difficulties in the quantitative analysis of Magnetic Resonance (MR) images. Thus intensity in homogeneities estimation is a necessary step before quantitative analysis of MR data can be undertaken. This paper proposes a new energy minimization framework for simultaneous estimation of the intensity in homogeneities and segmentation. The intensity in homogeneities is modeled as a linear combination of a set of basis functions, and parameterized by the coefficients of the basis functions. The energy function depends on the coefficients of the basis functions, the membership ratios and the centroids of the tissues in the image. Intensity in homogeneities estimation and image segmentation are simultaneously achieved by calculating the result of minimizing this energy. Furthermore, in order to improve its robustness to noise, the membership ratios are adapted by using nonlocal information. Experimental results on both real MR images and simulated MR data show that our method can obtain more accurate results when segmenting images with bias field and noise.