Abstract
Image segmentation is the process of assigning a label to every pixel in an image such that pixels with the same label share certain characteristics. The goal of segmentation is to simplify and change the representation of an image into something that is more meaningful and easier to analyze. Gaussian mixture model (GMM) and Student's-t mixture model (SMM) are well known and simple tools for image segmentation, but these methods suffer from several drawbacks. GMM and SMM have very poor results in the case of asymmetric distribution. In this paper, we present a new asymmetric mixture model for image segmentation. The proposed model uses Student's t-distribution that has heavier tails, multivariate asymmetric distribution and it is more robust than Gaussian. Simulation results demonstrate that our proposed method has a better performance than clustering methods which based on finite mixture models with probability density functions.