Program computes the parameter of GMM model using EM algorithm,optimized parameter thus obtained are used to calculate probability of every pixel with respect to number of classes(for given example it is taken as k).And than maximum probabilty of every pixel with respect to all classes is computed, on the basis of this pixels are labeled.
I did it long time ago if I remember correctly this is the paper:
Gaussian-Mixture-Model-Based Spatial Neighborhood Relationships for Pixel
Labeling Problem by Thanh Minh Nguyen and Q. M. Jonathan Wu
Thank you sir. Can you name reference paper for this implementation.
Code is optimized.
Download apps, toolboxes, and other File Exchange content using Add-On Explorer in MATLAB.