NVIDIA's AI turns regular videos into slow-motion videos.
There are also some advanced emulator software such as Twixtor's RE: Vision Effect that can create slow-motion videos smoothly but needs complex motion analysis in the clip and takes a lot of resources. time.
Nvidia used 11,000 standard slow motion videos of sports content with 240 native fps teaching an AI deep learning so that it could predict 210 frames before and after each frame.
Currently many high-end smartphones or cameras have been able to capture slowmotion clips of up to 240fps, but the resolution has been reduced. Nvidia's method is a much cheaper alternative to tens of thousands of dollars spent on cameras like Phantom. But compared to high-end cameras, Nividea's AI still needs time to process the video because all slow-motion video creation takes place after the video is recorded.
See more: