Microsoft's research paper introduces the AI model VASA-1, capable of turning a single image and audio clip of a person into a lifelike video of them lip-syncing.
The model surpassed technology from competitors like Runway and Nvidia, offering higher-quality and more realistic results.
VASA-1 can handle various inputs like artistic photos, singing audios, and non-English speech, showcasing its versatility.
The AI model was demonstrated to animate images in real-time with head movements and facial expressions, enabling quick generation of videos.
[
add
]
[
|
|
...
]