4 ms·
This sounds similar to k-means - pick some points "at random" and iteratively move them to the average location of all nearest neighbors until you stabilize. Bu
by web007 10y ago
This sounds similar to k-means - pick some points "at random" and iteratively move them to the average location of all nearest neighbors until you stabilize. But mean shift goes in the opposite direction, giving each point a window and grouping them on overlaps. Can someone who understands both confirm this is the case, or correct my understanding?
- RossBencina 10y agoThe article includes a comparison with k-means. From the article, mean-shift: > does not assume anything about number of clusters > can handle arbitrarily shaped clusters > is fairly robust to initializations > not very sensitive to outliers > time complexity mean-shift: O(Tn^2), k-means: O(knT). (k: number of clusters, n: number of points, T: number of iterations). [Edit: formatting]