More oddities with StarMask

[RBA]

When using scales 7 and above and Comp !=2 , the process goes to 98-99% then it sits there....
Forever and ever...
The program does NOT become unresponsive. I can abort the process, no problem.

The value of Comp may be circumstantial, but out of the few tests I ran (as you can imagine I didn't run all possible combinations!), when Comp=2 I always got an image, even with Scale=12 which of course it's the slowest of all.

[Juan Conejero]

I can't reproduce this behavior. I've tested with 12 scales and all values of the small-scale growth compensation parameter. Do you have this problem in a consistent way? Could you upload the image so I can reproduce it?

Anyway, All wavelet-based processes in PI, including StarMask, are going to have a major performance improvement very soon. Some tools will see execution times reduced up to 50%. Stay tuned I don't know if this is an efficiency-related issue though; it looks more like a memory issue.