3 ms·
Related: is there software that, given low-resolution video of people, can produce high-resolution images of their faces by analyzing multiple frames?
by lokl 7y ago
Related: is there software that, given low-resolution video of people, can produce high-resolution images of their faces by analyzing multiple frames?
- CommieBobDole 7y agoI'm not aware of anything that will do this offhand, but I think that's probably the future of upscaling and restoration - software that looks at a video, identifies all of the objects, learns as much as possible about their shapes, textures, lighting and movement by analyzing all the frames available, then makes an 'educated guess' about what it can't figure out and essentially renders a new scene in whatever resolution and framerate is requested. I can only imagine the amount of input data and processing power required to do that, though. Probably a decade or two away.
- acuozzo 7y agoWhat you're looking for is a temporal sharpening algorithm. These exist in crude forms because they're considerably more challenging to program than temporal noise reduction algorithms. Problem #1 is automatic image registration/alignment which is still largely an open problem of great interest to astronomers and those in medical/surgical fields. It's challenging because we're taking 2D images of objects moving in a 3D space. Problem #2 is classification. It's very hard at the moment to suss out which high frequency details are important vs. which are noise.