Alzheimer’s Disease Image Segmentation with Self-Organizing Map Network

Abstract

Detection and segmentation of Alzheimer's disease (AD) is very important because it provides structural information of abnormal and normal tissues. Functional Magnetic Resonance Imaging (FMRI) is a promising tool for detecting brain image. Alzheimer's disease (AD) FMRI image segmentation is treated as texture classification problem. In this paper proposed a novel method for AD image segmentation using Self-Organizing Map Network (SOMN). Four different types of features are extracted from the FMRI AD images such as the first order gray level parameters, multi-scale features, textural measures and moment invariant features. These features are used for image segmentation. Application of Adaptive Neuro-Fuzzy Inference System (ANFIS) model utilized for classification of Alzheimer's disease segmentation image. The proposed Segmentation and classification results are promising. Key word: Alzheimer's disease (AD), self-organizing map network (SOMN), ANFIS (artificial neural network fuzzy inference system), segmentation.