Nvidia's AI Tech Can Turn Standard Video Into A Slow-Motion Showcase

[kcarbotte]

Nvidia has developed a method using neural networks to convert standard video recordings into detailed slow-motion, ultra-high frame rate video clips.

Nvidia's AI Tech Can Turn Standard Video Into A Slow-Motion Showcase : Read more

 

[bit_user]

Motion interpolation is a hard problem that's pretty well-suited to AI. I wouldn't use it for slow-mo, so much as framerate conversion (i.e. converting between 24, 25, 30, 50, and 60 fps) and motion smoothing, like that featured in many TVs.

You could even use it to extrapolate frames, in VR or videogames, when frames are late leaving the rendering pipeline.

 

[Co BIY]

Pretty soon this tech will allow the easy "Photoshopping" of videos to include that which never happened.

You will not be able to believe your eyes.

 

[redgarl]

People are missing the point that this is not intended for any kind of video smoothing, it is intended for visual prediction.

Basically Nvidia AI car technology is having a huge issue. It is relating on video feed entirely. What Nvidia wants to do is introducing prediction capability for self-driving car. By analyzing cars behaviors feeds, they can predict how an autonomous car can react.

It is something necessary on the short term, but it doesn't solve the problem at all.

 

[pensive69]

'...By analyzing cars behaviors feeds, they can predict how an autonomous car can react....'

i'd add a suggestion to the auton vehicle types.
also remember they need to focus on what other human operated vehicles are doing and how they react.
the auton vehicle isn't out there alone.
maybe they also need a auton vehicle social network or neural network to communicate how they are reacting and manage any "bad actors".

 

[bit_user]

People are missing the point that this is not intended for any kind of video smoothing, it is intended for visual prediction.

Basically Nvidia AI car technology is having a huge issue. It is relating on video feed entirely. What Nvidia wants to do is introducing prediction capability for self-driving car. By analyzing cars behaviors feeds, they can predict how an autonomous car can react.

It is something necessary on the short term, but it doesn't solve the problem at all.

Except it's a method for video interpolation - not extrapolation.

And nowhere does the article, their blog post, or the research paper reference autonomous driving.

I'm not sure you realize just how many places AI is starting to show up. Nvidia is involved in lots of AI research that has nothing to do with autonomous driving.

BTW, if you're going to make factual-sounding posts from your imagination, it would probably be a good idea to start them with "I wonder if...". And if it is well-sourced information, then cite your sources.

 

[bit_user]

maybe they also need a auton vehicle social network ... to communicate how they are reacting and manage any "bad actors".

Yes, inter-vehicular communication networks are a real thing. Look it up.

 

[stdragon]

That's amazing! I wonder if this can already improve on the technique used in TVs to interpolate 30 FPS to 60 without creating the dreaded "soap opera" effect.

 

[bit_user]

That's amazing! I wonder if this can already improve on the technique used in TVs to interpolate 30 FPS to 60 without creating the dreaded "soap opera" effect.

Um, I think the "soap opera effect" is inherent in high-framerate video. It's entirely a subconscious association people have with the juddery 24 fps of film and the smooth 60i of soap operas. Leave your motion interpolater on, for a while, and the effect will naturally disappear. Then, you'll miss the motion interpolation whenever you don't have it.

https://en.wikipedia.org/wiki/Motion_interpolation#Soap_opera_effect

For me, the improvement in visual clarity of using motion interpolation is a huge win. The only negative is the artifacts, most of which Nvidia's solution could hopefully eliminate.

See also:

https://en.wikipedia.org/wiki/Display_motion_blur