Re:Text to Image Generation
by pugworthy on Sunday March 28, 2021 @11:02PM
(#26617272)
It would be interesting to assemble frames from a movie, say at scene changes, etc. and have this thing "narrate" a movie. It would be similar in concept (though not in content) to how narration for the visually impaired is offered as a form of assisted media viewing.
In fact, you could probably train it with existing visual descriptions from movies.
In fact, you could probably train it with existing visual descriptions from movies.