I have a problem with deciding algorithm for color quantization. The image that I want to do color quantization is an RGB image with resolution 512 x 512. I want to reduce the color value in the pixel to reduce the image size.
I don't want to use the popular algorithm k-means. And I found this mean shift algorithm.
Mean shift is a clustering algorithm (same as k-means) that have certain centroids and window (each centroids have 1) to look for densest part in its window then move the centroid into the densest part. It will keep going like that until it converge with another centroid and not moving anymore.
Is it possible to use mean shift in color quantization? Because both k-means and mean shift are clustering algorithms but all I found is image segmentation for mean algorithm and k-means for color quantization?