Nvidia AI creates realistic slow motion from standard video

Nvidia AI creates realistic slow motion from standard video

Technology News |
By Rich Pell

Their approach, say the researchers, outperforms a variety of state-of-the-art methods aimed at doing the same thing. With it, users will be able to slow down their video recordings after taking them rather than have to engage a slow-motion video mode – typically 240 fps – while recording.

“There are many memorable moments in your life that you might want to record with a camera in slow-motion because they are hard to see clearly with your eyes – the first time a baby walks, a difficult skateboard trick, a dog catching a ball,” say the researchers in a research paper on the project. “While it is possible to take 240-frame-per-second videos with a cell phone, recording everything at high frame rates is impractical, as it requires large memories and is power-intensive for mobile devices.”

The researchers used Nvidia Tesla V100 GPUs and a cuDNN-accelerated PyTorch deep learning framework to train their system on over 11,000 videos of everyday and sports-related activities shot at 240 fps. Once trained, the convolutional neural network predicted the extra frames needed to make videos shot at a lower frame rate look more fluid and less blurry when slowed down.

“Our method can generate multiple intermediate frames that are spatially and temporally coherent,” say the researchers. “Our multi-frame approach consistently outperforms state-of-the-art single-frame methods.”

To demonstrate their method, the researchers using a series of clips from The Slow Mo Guys – a popular slow-motion-based science and technology entertainment YouTube series – and made their videos even slower. For more, see “Super SloMo: High Quality Estimation of Multiple Intermediate Frames for Video Interpolation.”


Related articles:
AI lip-sync algorithm puts words in someone’s mouth
Amazon deep learning video camera now shipping
AI video analysis software adds auto-rotated models, Linux support, API

If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :    eeNews on Google News


Linked Articles