New neural network creates an animated 3D face model based on voice
The standard animation uses the Digital Mark character model, but it can be changed to a face with any other features or even an animal. The program can generate emotions on a 3D model of a person's head in real time. All you need to do to get an animation is to upload a voice recording. Depending on the language and intonation, the standard model changes the facial expression and articulates everything in the audio recording. The program runs on the open platform Nvidia Omniverse. There you can make a model in real time.  The program is now available in English, French, Italian and Russian. 
2021-11-12