Open thesamovar opened 9 years ago
For example, start with the mask that has the smallest number of points. Find the closest mask to it. Merge these two. Repeat until you have the number of masks needed for mask starts. @shabnamkadir what do you think?
Could improve quality of initial clustering?