I actually quite like starless images. I posted an image processed with this a little while ago now but didn't gather much interest, probably because everyone here seems to dislike such images.
According to the github repo:
Quote:
|
This newral net was trained using data from refractor telescope (FSQ106 + QSI 683 wsg-8), so will work best for data from similar imaging systems. That's a bit of a bad news for many users of reflector telescopes. If you have long spikes in your images, then the net will not take care of these spikes very well, I tried and didn't like results too much.
|
Said github repo seems to show examples of images with diffraction spikes that have turned out well... so now I'm confused. I can't remember if it had those examples in the past... it . definitely had some, but I'm not sure about the ones with obvious diffraction spikes. Maybe the description is outdated now and the author has successfully trained it with such images.
I process my images with unmasked stars so they probably come out a bit worse than they could. I've recently started playing around with some ideas to clean those up, but maybe I should revisit this software. Seems like this could be an easy way to get rid of them so I can replace them with stars that have been better processed.