Max, not sure if this suffices. If you view our free primer tutorials http://www.ip4ap.com/pixinsight_part1.htm you'll note the ability to create a new instance of any process, and to save it for later recall. You could experiment with your data, and when happy with the baseline- save it as a new instance icon. When you purchase our full tutorials, you get my personalized icon set with starting points, which in some cases may be more in line with your data than the actual tool defaults. FWIW, I like MultiscaleLinearTransform for linear, and TGVDenoise for nonlinear noise reduction. You'll also see great settings in Juan's tutorials here on the forum.
Warhen,
I was hoping there are some unifying mathematical concepts and standards that Juan and other use in developing their modules.
For example , your own tutorial on TVG notes that a good starting for the edge setting is the standard deviation of the image background .
What other goodies are lurking beneath the parameter settings that have a statistical basis that could be exploited ?
We have a script that separates images layers.
Could we created a script that measures the noise in each and suggests ranges for mulltiscale noise reduction etc.?
Just thinking out loud...
Max
Max
