Add a mask for the QSpace conversion
This MR adds an input for a mask image to discard some pixels for the QSpace conversion. Pixels with non-zero value in the mask are discarded (Is this convention OK?). The mask is expected to have the size of the images.
The mask is handled during the computation of the histogram look-up table.
As it is now, it is mutually exclusive with the use of binning and median filter. I can make the mask and image binning working together if needed, but I see little interest in it as the next step is a histogram. Let me know if you want it.
Merge request reports
Activity
Related to #13
Dear Thomas,
I think the binning is not very useful and if it does not work with the mask, it is fine but it should be clear to the user. I will merge and test.
I think both the binning and median_filter are compatible with the mask. Also I have some scripts which combine these. So I think making it work would not be too much work. But I also have the impression that the binning is not very useful in general and could be removed completely?
mentioned in commit c73a2254
Hi, Yes we can combine mask and binning and median filter. I though you used the median filter only to get rid of 'hot pixels' and that the mask would be a better option to do this. Let's just support both mask and median filter, it shouldn't be a big deal.
If the binning is no more useful, I would be in favor of removing it.
mentioned in issue #39 (closed)
mentioned in issue #40 (closed)