A fast, vectorised implementation of the K-Means clustering algorithm intended for use with image clustering.
The vectorised implementation of the algorithm is contained within KMeansRGB.m. This algorithm needs to be supplied with starting points which are generally randomly sampled from within the image colour sample space. An implementation of this is within SelectKRandomPoints.m. To generate an image from the clustered points, pass the output to CreateKColourImage.m.