Image clustering by Gaussian mixture models

I want to use GMM(Gaussian mixture models for clustering a binary image and also want to plot the cluster centroids on the binary image itself.

I am using this as my reference: http://in.mathworks.com/help/stats/gaussian-mixture-models.html

This is my initial code

     I=im2double(imread('sil10001.pbm'));
    K = I(:);
    mu=mean(K);
     sigma=std(K);
     P=normpdf(K, mu, sigma);
     Z = norminv(P,mu,sigma);
     X = mvnrnd(mu,sigma,200);

     scatter(X(:,1),X(:,1),10,'ko');

    options = statset('Display','final');
    gm = fitgmdist(X,1,'Options',options);

     idx = cluster(gm,X);
     cluster1 = (idx == 1);
      cluster2 = (idx == 2);

    figure;
     scatter(X(cluster1,1),X(cluster1,1),10,'r+');
     hold on

      scatter(X(cluster2,1),X(cluster2,1),10,'bo'); 
     hold off
     legend('Cluster 1','Cluster 2','Location','NW')

But the problem is that the output is coming like a straight line and not like the figures which are mentioned in the mat-lab reference .And also I am unable to plot the cluster centroids received from GMM in the primary binary image.How do i do this?

I want the output to be like this The contours represent the position of the cluster centroids

Answer 1

[missagh dayer]

Hi, Newman. I ran your code on my own computer and detected a couple of problems:

Problem #1: You used this line:

gm = fitgmdist(X,1,'Options',options);

From what I gather you want to perform clustering on an image, so why are you performing it on X? X is a randomly created vector... You need to perform the clustering on P! So your code would look like this:

gm = fitgmdist( P,1,'Options',options);

(P.S why are using only 1 cluster in you code? Unless you use 2 or more, this code will have no effect!)

Problem #2: Your input was a 2-D image, so you need to convert your clusters into 2-D images as well. To do that you might want to use something like this:

 cluster1 = (idx == 1);
 cluster1=reshape(cluster1,size(I));
 imshow(cluster1);

Do this for all clusters and see what you get.

As for the centroids, I don't know what code you are using so I'm afraid I can't be of much help there.