THIS IS K MEANS CLUSTERING SEGMENTATION CODE k-means algorithm proposed by Duda and Hart: A. Initialise the position of the means m1 → mk. B. Assign each of the k-items to the cluster whose mean is nearest. C. Recalculate the mean for the cluster gaining the new item and the mean for the cluster loosing the same item. Recalculation is made using the variance. D. Loop through steps (A) and (B) until there are no movements of items. YOU CAN USE IMAGE FROM http://www.cse.yorku.ca/~mridataset/ to check it