- K means algorithm is performed with different initial centroids in order to get the best clustering.
- The total cost is calculated by summing the distance of each point to its cluster centre and then summing over all the clusters.
- Based on the minimum overall cost achieved during each iteration of 'iterKMeans' the pixel assignment to their respective clusters are made and final compressed image is obtained.
- This algorithm will run slower as the number of clusters , size of the image and number of iterations increase.