Skip to content

Imitator

Personalized speech-driven 3D facial animation

Visit official site

What Imitator is

The method is described as a personalized speech-driven 3D facial animation approach that generates subject-specific motion sequences.

What you can do with it

  • Drive 3D facial animation from an audio sequence
  • Produce motion with accurate lip closure
  • Compute a style embedding from a short reference clip

Application scenarios

  • Give movement to characters inside games
  • Use the method in virtual reality settings
  • Apply it within remote meeting systems

Pros & cons

✓ What we like

  • Accurate lip closure for bilabial consonants m, b and p
  • Style embedding can come from a five-second reference video

! What to watch out for

  • A short reference video is used for the style embedding
  • The output is a subject-specific motion sequence

FAQ

What are the inputs?

An audio sequence and a personalized style embedding.

Where does the style embedding come from?

The material states it can be computed from a short reference video such as five seconds.

Last reviewed: 2026-11-15

More AI video generation tools

View all →

How we review