mg262
6th November 2005, 08:36
I've recently run across a few video algorithms that learn based on training material... e.g. deinterlacers that require a large amount of progressive training material. Some of these wouldn't be too hard to code up for AVISynth, but at the moment I can't think of any sensible way to get hold of a large body of training material. (Obviously the filter will only really work on material similar to what is trained with, so e.g. just training it on animation isn't a great idea.)
Anyone have any ideas...?
Anyone have any ideas...?