How to cluster the feature vectors(which have 8 feature as image in cell ) using fuzzy c means clustering?How to map with the original image after clustering?

I am using 'Madurai satellite image' as the dataset. My objective is to categorize the pixels as urban,water,rural,building etc..First, the input image is applied to laplacian pyramid filter and I had taken HH band for further processing. HH band is further applied to directional filter bank with level=3 to get 8 features as image.I don't know how to apply these feature vectors( which have 8 feature as image in a cell) to fuzzy c-means clustering? And how to map the clustered result with the original image to get the final result? Kindly help me sir.

Answer 1

[KSSV]

Read about kmeans ..this is very easy to use. Let I be your image and you want to cluster them into 8 groups.....check the below pseudo code.

I = rand(100) ;
k = 8 ; % groups 
idx = kmeans(I(:),k) ;

figure
hold on
for i = 1:k
    plot(I(idx==i),'.','color',rand(1,3))
end