How do I go about running k-means on my grayscale video sequence?

Question

I have a sequence of 2D grayscale images for which I would like to run the k-means clustering algorithm on. My data is in the form m x n x p where p is the number of frames.

The matlab code available in the statistics toolbox expects the data X to be in the form n-by-p data matrix where the "Rows of X correspond to points, columns correspond to variables."

Can I use this code? How do I adapt my data to be in the expected format?

Thanks in Advance,

n

p.s. Also, if possible I'd prefer not to use a function from the statistics toolbox, if somebody can recommend something else (such as a mex function). It would be much appreciated.

Answer 1

You can try to use this file exchange function.

It the otsu algorithm base on multi-class segementation.

For multiclass, you can use im2bw