Here is my workflow for one of my older DSLR’s that has quite a few hot pixels:
-Bring photos (.NEF) into lightroom classic (at this point, hot pixels aren’t visible, LRC seems to be correcting for them)
-File → Plug in extras → Process with topaz photo AI
-Usually noise reduction and sharpening (wildlife photography)
-Process photo in LRC (now the hot pixels are visible and LRC doesn’t seem to be correcting for them in the .DNG from topaz)
It’d be nice if topaz photo AI could fix the hot pixels so that I don’t need to fix them manually.
Given the pixels are almost certainly in the same place every time I would think this should be possible assuming LR passes on enough info about the camera that PhotoAI could offer specific training for these anomalies and remove them in the process. While I would hope the AI could eventually tell what a hot pixel is it is probably easier to train it locally by telling it these spots are not meant to be there. The issue is if you were using it on the night sky some of those hot pixels may look similar to stars or planets and you might not want to lose those.
That’s precisely the difficulty. It’s the same idea with fine details being mistaken for noise and vice versa. Our developers are currently investigating this though!