One-Class Support Vector Machines Based Cluster Validity in the Segmentation of Hyperspectral Images

Abstract

In this paper, a novel cluster validation method based on one-class support vector machines (OC-SVM) is presented. Also it is proposed to segment hyperspectral images with subtractive clustering accompanied by phase correlation. The proposed cluster validity measure is based on the power of spectral discrimination (PWSD) measure and utilizes the advantage of the inherited cluster contour definition feature of OC-SVM. Basically this method provides a solution to the estimation of the correct number of clusters which is an important problem in hyperspectral image segmentation.