Image segmentation using k means

Hi All I have a feature vector of an image now i want to segment the image using k-means clustering algo. can any one provide me the code..

[Walter Roberson]

Please see http://www.mathworks.com/matlabcentral/answers/43073#answer_52939

[M@lik Ali]

Sorry there is no answer....

[Walter Roberson]

There is, however, remarks about what it means when your Question is tagged with "doit4me".

[mahesh]

hi wat feature did u extract from the image??

Answer 1

[Image Analyst]

Here's an example for you:

Color-Based Segmentation Using K-Means Clustering

http://www.mathworks.com/products/demos/image/color_seg_k/ipexhistology.html

[M@lik Ali]

Thanks to reply, but in my case i have a feature vector of 6 rows and 6114 columns. so how i can use this code, please give me the solid solution.

[Image Analyst]

I guess you have 6 image analysis measurements for each of 6114 blobs in the image? I can't give you the solid solution - I don't have the Statistics toolbox. Perhaps someone who uses kmeans can help you.

[M@lik Ali]

this code return this error also ??? Error using ==> kmeans Too many input arguments.

Error in ==> KmeansI at 21 [cluster_idx cluster_center] = kmeans(ab,nColors,'distance','sqEuclidean', ...

[Anton Semechko]

I have an implementation of bisecting k-means algorithm here:

http://www.mathworks.com/matlabcentral/fileexchange/34807-nonlinear-relaxation-labeling-for-image-processing

the function is called 'MyKmeans'

[Walter Roberson]

Which MATLAB version are you using? The demo probably assumes you have a very recent version.

[M@lik Ali]

version 7.10 win 32

[Walter Roberson]

At the command line, please command

which -all kmeans

and tell us the result.

[M@lik Ali]

Ohuh thanks, there ware two function with same name now its working, but now it return me the feature vector like matrix size is 8X6

0.1000 1.0428 0.6896 0.0012 0.0014 0.0001 0.1000 1.0428 0.6896 0.2068 0.1074 0.0305 0.1000 1.0428 0.6896 0.0232 0.0644 0.0025 0.0710 0.7404 0.4896 0.0359 0.0356 0.0031 0.1302 1.3572 0.8975 0.0057 0.0029 0.0002 0 0 0 0.0027 0.0063 0.0004 0.1302 1.3572 0.8975 0.1287 0.1909 0.0496 0.1151 1.2000 0.7935 0.3897 0.3009 0.1020

So is it segmented image if yes then how many regions it have and how i can color the different regions actually the feature vector i give input to the kmeans was LUV components..