Separation of brain tissues in MRI based on multi-dimensional FCM and spatial information

Abstract

Due to intensity non-uniformity (INU) and noise brain magnetic resonance image (MRI) segmentation is a complicated concern. Many methods have been presented to overcome brain MRI segmentation. Among these methods, using fuzzy c-means (FCM) is introduced as an effective strategy. Spatial information cannot be considered at a standard FCM. Therefore, many methods have been presented to optimize standard FCM with optimization of objective function. In this research work, a novel method has been proposed for brain MRI segmentation (BMS) based on multi-dimensional standard FCM. In this technique, different features of neighboring pixels such as mean, standard deviation and singular value in combination with pixel intensity has been used for typical pixel segmentation. The results have been evaluated against manual segmentation on a publicly available dataset.