Bayesian Framework for image segmentation based on Nonparametric Clustering with Spatial Neighborhood Information
Imane Kirati, Yamina Tlili · International Journal of Computer Applications · 2011
In this paper, we present a Bayesian framework for image segmentation based upon spatial nonparametric clustering. To estimate the density function on a nonparametric form, the proposed model exploits local Gaussian kernels. In addition, we have incorporated the spatial information to the clustering process by adding a spatial function for weighting the posterior probabilities.The main advantages of this model are two. First due to the non parametric structure, it does not require the image regions to have a particular type of density distribution. Second, adding spatial information yields more homogenous and smoothed regions.The experimental results based on real images demonstrate the efficiency of the proposed method and indicate clearly its robustness to noise. General Terms Image segmentation, image processing,pattern Recognitioncomputer vision.