Hi Phillippe
Yes, I know I've been away from these tools for a while. But, I have not been away from development of TGV related tools at all. In the past months I've been doing some research on automatic calculation of the optimized parameters in TGV problems. Now I have a couple of predictors that are correlated with lowest mean square errors in the reconstructions (or noise reduction) and also have the functions to calculate the functional costs to build L-curves (
https://www.sintef.no/globalassets/project/evitameeting/2005/lcurve.pdf). I plan to introduce both estimators in the next release.
Also, I performed some exhaustive research on the error maps, varying both the strenght and edge protection parameters. I think that I have a way to simplify the implementations, to make them more user friendly. In a nutshell, I found a way to use a change of variables that allows the user to explore the changes in the direction of greater impact on the result, while the other direction has the lesser impact. So, we may merge the parameters into one critical new parameter, and a couple of "advanced parameters" that I plan to make accesible in a cleaner way for fine tunning.
On other fronts, not that much related with our astronomical images, I have expanded the functions to work with 3D data (data cubes). This is useful for biomedical images mainly. And, I have builded a TGV-QSM function that helps to find magnetic susceptibility measurements in the brain. I know, quite a lot off-topic here, but this is my main research front at the time.
Anyway, I'm trying to reassign some time to development of tools for PI. I have to attend a big conference next month in Singapure, so I'll try to have something new in the next weeks before leaving.
