2024-04-18 10:56:23
Microsoft has unveiled artificial intelligence that allows you to create a video from a photograph.
“The first AI-generated video that looks super real.” In a message shared on
“Hyper realistic” videos
This AI, called VASA-1, uses a photo taken in portrait mode with sound featuring the guinea pig’s voice, to produce a “hyper realistic” video, which allows for precise lip synchronization, realistic facial animations and natural head movements.
The result, visible below, is particularly stunning: “It’s incredible, given that the AI-generated video looks so real,” says the manager.
But as she points out, this video is only an example given in a compartmentalized environment to maximize results. However, this is a great feat for Microsoft, which continues to spend billions each year, particularly with OpenAI, to position itself as a key player in the field of artificial intelligence.
For the moment, VASA-1 remains quite limited, in particular because it can only generate videos in 512×512 megapixels, and in 40 frames per second. However, in the future, this AI could significantly improve the creation of realistic avatars, for example in video games, but also in the context of educational tools. However, it is not possible at this stage to create an entire body, while the hair is not animated either.
All of the demos distributed by Microsoft are visible on the official blog of the project. We see very real people, but also paintings or concepts that come to life with significant ethnic diversity.
At the moment, this AI is not available to the general public.
Most read
1713439825
#Microsoft #unveils #stunning #transforms #simple #photo #video