AI applied to videoconferencing

Maxine. © Nvidia.

Videoconferencing systems have become an overnight must for many working stiffs. But who hasn’t experienced some or all of the following problems with them: bandwidth slumps that freeze the image, distracting background noises, or just plain poor video. To address the most common videoconferencing issues, Nvidia launched a new videoconferencing toolkit targeted at developers, called Maxine. The system processes video data in the cloud, on Nvidia GPUs. Using AI, Maxine can realign callers’ faces and gazes so that they’re always looking directly at their camera, significantly reduce the bandwidth requirement for video by only transmitting “key facial points,” and upscale the resolution of videos. Other features available in Maxine include face re-lighting, real-time translation and transcription, and animated avatars.

⇨ YouTube, “AI-Powered Video Conferencing with NVIDIA Maxine.”

⇨ YouTube, “Inventing Virtual Meetings of Tomorrow with NVIDIA AI Research.”

⇨ The Verge, James Vincent, “Nvidia says its AI can fix some of the biggest problems in video calls.”