Automatic MRI Image Threshold using Fuzzy Support Vector Machines

Analysis and segmentation of the basal ganglia of the brain is a major issue in brain MRI image processing. In areas such as segmentation, calculate the shape of the surface,texture and its analysis forms the basis of the methods. In this study, texture classification is selected and studied as one of themost important subject matter in the field of tissue analysis. To classify ganglions first features must be extracted from the imageproperties and then classification algorithm applies on them. There are several methods for feature extraction and selection that their goal is providing data that are suitable forclassification. In this paper, we used a combined method for feature extraction. Then to classify ganglia the SVM methodused. In this article, a similarity measure based on the Euclidean distance defines fuzzy membership function. SVM classification performance is improved with this fuzzification method. The results of experiments that compared with other methods show that the proposed method is an accurate method.

Keywords:

Image Threshold , Image segmentation , Fuzzy Support Vector Machine , classification , Fuzzy membership function