
NVIDIA introduced a platform for the developers of Nvidia Maxine, which is designed to correct the main disadvantages of video calls. Details about the project are given by The Verge .
The Maxine platform allows video calls using NVIDIA cloud services that process and improve video using artificial intelligence. In particular, algorithms allow you to change the position of the interlocutor’s head in such a way that his gaze is directed to the camera while maintaining facial expressions. Also, the functionality of the platform allows you to reduce the quality of the connection for video calls - the system allows you to capture "key points on the face", and then processes the video, improving its quality before transferring to the device of the interlocutor.
The demonstration video also shows that Nvidia Maxine supports speech recognition and translation from a foreign language in real time, makes it possible to create animated avatars and use high -quality backgrounds. Finally, algorithms can drown out a background noise that prevents interlocutors from communicating.
The Verge notes that the listed functions are not fundamentally new, and some of them are already used by others in their own video call services, but in NVIDIA they are counting on the success of the MAXine platform. For this, the services of the platforms should use third -party companies for which Maxine was created. So far, NVIDIA has not reported partners who were ready to introduce Maxine into their video call services, saying only that the company is negotiating with a number of potential customers.
Recall that in mid -September, NVIDIA officially announced the purchase from the Japanese SoftBank of the British developer of chips for ARM mobile devices. After a transaction worth almost $ 40 billion, which the regulators must first approve, ARM will continue to work as a NVIDIA unit, and the company's headquarters will still be located in the UK. ARM will also continue to work as part of a license model, allowing manufacturers of smartphones and other devices, including Apple, Samsung and Qualcomm, to produce chips with ARM architecture.