display output k-means clustering, display output clustering as a image

Hello,

I have a image, name image :test 3

I,map]=imread('test3','bmp');
I = ~I;

imshow(I,map);

[m n]=size(I)

P = [];
for i=1:m    
    for j=1:n        
        if I(i,j)==1
            P = [P ; i j];        
        end
    end
end

size(P)
MON=P;

[IDX,ctrs] = kmeans(MON,3)

as I plot the clusters in the image, resulting

I want to draw idx and ctrs in the image.

I don't know, How do I get back image with 3 new cluster(each cluster, different color in the image)

can anyone help ?

Thanks.

[Salaheddin Hosseinzadeh]

Image Analyst should be able to help!

[Image Analyst]

Sorry, I don't have the stats toolbox, which is what has kmeans.

[Tomas]

I just want to know, how to convert ouput save in cell, back to image. Thanks.

Answer 1

[Dishant Arora]

[ I map] = imread('test3.bmp');
I = ~I;
imshow(I,map);
[m n]=size(I)
P = [];
for i=1:m    
    for j=1:n        
        if I(i,j)==1
            P = [P ; i j];        
        end
    end
end
size(P)
MON=P;
[IDX,ctrs] = kmeans(MON,3);
 clusterImage = zeros(size(I));
 clusteredImage(sub2ind(size(I) , P(:,1) , P(:,2)))=IDX;
 imshow(label2rgb(clusteredImage))

[Tomas]

Warning: Image is too big to fit on screen; displaying at 0% > In imuitools\private\initSize at 73 In imshow at 262

does not display the image,

[Dishant Arora]

Just some typing error, It should be:

clusteredImage = zeros(size(I))

[Tomas]

Ok, thanks, I would like to ask, when i have coordinates the image save in cell, how i display cluster in the image.

I enclose my matlab code.

MAIN FEATURES name metodakkk.

Thank you very much

[Dishant Arora]

You want to show individual clusters in separate images?? if yes:

Z = cellfun(@(Z) Z', Z,'Un',0);
Z = cellfun(@(Z) cell2mat(Z), Z,'Un',0);
Z = cellfun(@(Z) sub2ind(size(I) , Z(:,1) , Z(:,2)), Z,'Un',0);
for ii = 1:length(Z)
    figure
    cluster{ii} = zeros(size(I));
    cluster{ii}(Z{ii}) = 1;
    imshow(cluster{ii})
end

[Tomas]

I want show all cluster in one image.

Thanks.

[Dishant Arora]

Z = cellfun(@(Z) Z', Z,'Un',0);
Z = cellfun(@(Z) cell2mat(Z), Z,'Un',0);
Z = cellfun(@(Z) sub2ind(size(I) , Z(:,1) , Z(:,2)), Z,'Un',0);
figure
cluster = zeros(size(I));
for ii = 1:length(Z)

      cluster(Z{ii}) = ii;

end
imshow((cluster))

[Tomas]

one more question as a color different clusters ?

Thank you very much for your help

[Dishant Arora]

imshow(label2rgb(cluster))

[Tomas]

Thank you very much for your help

Answer 2

[rizwan]

Hi Experts, I am using the following code to find clusters in my image using K - Mean [ I map] = imread('D:\MS\Research\Classification Model\Research Implementation\EnhancedImage\ROIImage.jpeg'); I = ~I; imshow(I,map); [m n]=size(I) P = []; for i=1:m for j=1:n if I(i,j)==1 P = [P ; i j]; end end end size(P) MON=P; [IDX,ctrs] = kmeans(MON,3,'display', 'iter','MaxIter',500); clusterImage = zeros(size(I)); clusteredImage(sub2ind(size(I) , P(:,1) , P(:,2)))=IDX; imshow(label2rgb(clusteredImage))

The out put of the above code is

>> ImageEnhancement

m =

   180

n =

   317

ans =

       20306           2

iter   phase       num           sum
   1       1     20306   9.40619e+07
   2       1      2727   7.34318e+07
   3       1       876    7.1216e+07
   4       1       574   7.03212e+07
   5       1       410   6.98473e+07
   6       1       298   6.96024e+07
   7       1       173   6.95038e+07
   8       1       122   6.94633e+07
   9       1        65     6.945e+07
  10       1        45    6.9445e+07
  11       1        30    6.9443e+07
  12       1        15   6.94424e+07
  13       1         8   6.94422e+07
  14       1         3   6.94422e+07
  15       1         1   6.94422e+07
  16       2         0   6.94422e+07
Best total sum of distances = 6.94422e+07
Warning: Image is too big to fit on screen; displaying at 2%

Can any one explain this out put and how can i see proper out put of K- Mean???

I shall remain thank full You To

Regards

Answer 3

[Yanyu Liang]

It shows how the kmeans is going at each iteration. "kmeans" implementation in matlab has two phases (you can think of it as two different approach to update assignment), so "phase" just tells if it is using first phase or second. "num" tells the number of points that change their assignment at that iteration (as you can see when it hits zero, the algorithm stops). "sum" is the objective value "kmeans" is trying to minimize.