Jensen Huang, the boss of Nvidia, willingly lends himself to his company’s experiments to demonstrate the superiority of its technologies. New example with this 3D bobble head version, which looks like a connected assistant!
After an extremely realistic 3D version which amazed everyone, the CEO of Nvidia once again served as a model to demonstrate a technology of the company. The latter has developed a reference application for the Tokkio “talking kiosk” project, and Jensen Huang is the guinea pig.
Impressive demonstrations of Nvidia technologies
The app has created a 3D avatar of Jensen Huang that looks like a toy or those big-headed figures. In addition to the graphic representation and “stop motion” animations, Nvidia has injected artificial intelligence into the character. As a result, he was able to answer questions about climate change and the role of proteins in the human body.
Nvidia recently unveiled several tools for creating avatars, realistic or not, under the banner of the Omniverse platform. It mainly consists of two parts: Omniverse Avatar for the creation of characters, and Riva for the management of the voice. By combining these technologies, we obtain results that are worth seeing, and which could represent the future of connected assistants.
Nvidia’s talking and intelligent avatars offer to “chat” not just with simple robotic voices, but also with human representations. In the demo below, Nvidia used the Maxine Project Toolkit to create an avatar from a photo. Riva takes care of the polyglot voice (Chinese, German, English, French…), with intonations which reinforce realism.
Developers can get their hands on Maxine already, while Riva is available in a public beta. Everything is free for “small” projects, but a Riva Enterprise program will be released next year. Part of the Omniverse platform is currently in public beta, but Omniverse Avatar is still in development.