I came back to my image of the Deer Lick Group and Stephan’s Quintet from last year due to the JWST reveal. While my original take on the image was not terrible for the tools and knowledge I had available at the time, I wanted to give the original integration another go. The new neural network-based denoising algorithm is a huge improvement over the TGV and multiscale denoising I used to use. Starnet V2 also let me remove the stars mostly pretty accurately so that I could correct dust motes and apply a morphological transform to reduce the stars a bit. I was also more careful with color calibration and saturation this time. Except for a few defects from the flats, this data was really quite good. I think I did a little better job this time around,
