Nvidia is using AI to improve video calls and conferences



(Pocket-lint) – Nvidia has announced a new platform known as Maxine that’s designed to reduce the bandwidth of your video calls while also improving quality. 

As you’d expect, Nvidia is using artificial intelligence in order to enhance the video quality while reducing the impact on bandwidth. Nvidia Maxine uses AI to analyse “key facial points” of people on the call and then to reanimate their faces to reduce bandwidth requirements while passing the data through. 

Nvidia claims that this technology uses as little as one-tenth of the bandwidth of a standard video call. It also runs in the cloud, thereby not using processing power on your computer either. 

It’s interesting seeing Nvidia Maxine at work, essentially using AI to recreate your face for others to see, then animating the picture as it thinks it should. In some ways, this is similar to the way Deep Fake videos work. The AI is intelligent enough to handle users wearing hats, masks and glasses. 

Nvidia has also shown off other ways the AI could be useful, like re-aligning the view of your camera so it looks like you’re looking at the people you’re talking to, even if you’ve turned to look at another screen. 


Maxine can even be used to transform people on the call into animated characters instead. Using specific points from the original video to add clothing, hair, accessories and more and give you an entirely new look.

Alongside all the AI intelligence, Nvidia Maxine also brings other improvements to the video calling experience including intelligent noise cancellation, real-time audio translation and transcription.

Nvidia Maxine is being made available to developers in order to improve current video conferencing tools and future ones. The result should be a better experience for everyone, even those with slow broadband and lesser quality webcams. 

Writing by Adrian Willings.


Source link