Can Video Enhance AI sharpen video?

Is there a way to sharpen video (like sharpen AI does for images) without changing resolution?
Right now I use sharpen AI to sharpen video that is saved as image sequences, then put those images back into a video. There should be an easier way.

I use Photoshop to break video frames into jpegs, batch the jpegs via PS, then use VirtualDub to stitch back into a video.

1 Like

Sure. If you use the Proteus model (or Theia, for that matter), there are separate sliders for both sharpening and detail recovery. Just experiment with those while setting the output resolution to 100%. There are also sharpening algorithms in freeware like DaVinci Resolve which are pretty good.


I forgot to say the thing I want to sharpen is an out of focus talking head (it’s an interview video) so I need some sort of AI face sharpen.

You don’t have to change resolution to use ANY model.

I usually use the first pass for cleaning up the video and use the second for upscaling. I have to use two-pass for most 80s and 90s TV on DVD.

Why in 2 steps? Do you use different AI models for cleaning and upscaling?

I often need two steps because the source material was just put on DVDs.

It wasn’t necessarily cleaned up.

The first model I use depends on the material, but I always finish with Artemis HQ for the upscaling to 1080p.

The only DVDs that I am upscaling that don’t need serious work are my super-bit version of “The Fifth Element” and (though I haven’t done then yet) the “Back to the Future” Trilogy.