Applying k means to matrix

I have the result after dividing same pixels into blocks,

result is in variable out in workspace

27  [1,9]
30  [1,12]
31  [1,13]
32  [1,14]
33  [1,15]
34  [1,16]
35  [1,17]
36  [1,18]
37  [1,19]
38  [1,20]
39  [1,21]

now i have to apply k means for the above result ,i.e for each block i have to apply k means 

please help

[Walter Roberson]

Okay, go ahead. The function name is kmeans()

[Pat]

walter i tried 

z=out(:,2);
 z2=cell2mat(z);
 for i=1:length(z2)
f=kmeans(2,z2(i));
end

i get error 

Error using ==> kmeans at 283
X must have more rows than the number of
clusters.

that y i posted these question since i got error

[Walter Roberson]

You are overwriting "f" on each iteration of the loop.

What do the arguments "2" and "z2(i)" represent to you?

[Pat]

walter i want to apply for all rows,in this 1st column represents pixel value and next column represents pixel position

now please tell how to apply k means for each block,i dont have idea for applying to each block the previous code i used for separating into blocks is

S is my image of 256x256;

   d = unique(S);
  out = [];
  for jj = 1:numel(d)
      s=regionprops(bwconncomp(S == d(jj),4),'PixelList');
      out = [out;[repmat({d(jj)},numel(s),1) {s.PixelList}']];
  end
  [ii,ii]=sortrows(cell2mat(cellfun(@(x)x(1,:),out(:,2),'un',0)),[2 1]);
  out = out(ii,:);

now for out i have to apply kmeans,i.e for each block

[Pat]

please look at this

http://www.sendspace.com/file/6udp7y

my output is like that ,i have to apply kmeans for each block,please assist

[Walter Roberson]

I don't understand what that code is intended to do, and I do not understand how you ended up (in your file) with multiple entries with the same pixel value.

Anyhow, let us stick for the moment with the question of what kmeans(2,z2(i)) means to you. Why is the first argument always 2? Why is the second argument variable? Which argument represents the number of clusters? Which argument represents the data to be clustered?

[Pat]

z2(i) represents each row,applying k as 2,second argument is data ,so it is every row which i represents,number of clusters is 2

[Walter Roberson]

Time for you to re-read the documentation for kmeans to see what arguments it expects to be passed in.

[Pat]

Walter i have read the documentation and have understood ,but no idea how to implement after diniding into blocks ,I know to apply for image or matrix ,but to the output below i dont know how to apply

http://www.sendspace.com/file/6udp7y

[Walter Roberson]

Okay, so what is your current version of the code (after taking the documentation into account) ?

[Pat]

    Walter i tried 

    assuminug my values  are 

    S=27  [1,9]
    30  [1,12]
    31  [1,13]
    32  [1,14]
    33  [1,15]
    34  [1,16]
    35  [1,17]
    36  [1,18]
    37  [1,19]
    38  [1,20]
    39  [1,21]

    S=S(:,2);
    S=cell2mat(S);
    for i=1:length(S)
    D=kmeans(S(i(:)),2);
    r=reshape(D,1,[]);
    end

dont know whether it is correct or not ,and i saw a reference code in 

http://www.mathworks.in/matlabcentral/answers/46705-how-images-are-divided-into-sub-images

here the peppers image is divided into many parts,like that i want ti divide image ,such the each subimage has same pixel value,and then apply k means for each block