SenseAvatar
Generates lip-synced digital-human video from a portrait and a voice track
WHAT IT IS
SenseAvatar is a digital-human video capability SenseTime introduced in August 2022 and offers as an interface. Given a portrait and a voice track, it generates a talking digital human whose lip movement matches the speech, with natural head motion. It is used for virtual presenters, announcements and customer service. The capability is a closed service.
Why it matters
It reduced digital-human generation to “one photo plus one voice track”, letting lip-synced avatar videos be produced at scale through an interface — an early example of digital humans moving toward large-scale use.
Key specs
- Input
- A single portrait and a voice track
- Output
- Lip-synced talking digital-human video
- Availability
- API
- Open weights
- No
Capabilities
Related concepts
Autoencoders & VAE
Squeeze information through a bottleneck, then let it grow back
Generative Models: An Overview
Discriminative models answer "what is this"; generative models answer "what should this look like"
Multimodal Generation
One model that learns to speak, to draw, to move — even to model the 3D world