2024-04-21 04:52:11
Microsoft Research Asia has released a white paper on the AI applications the company is developing. The program is called VASA-1 and can create very realistic videos (Deepfakes) from a single image of a face and background music. Not only that, the software can create videos and swap faces in real time.
Visual Affective Skills Animator or VASA is a machine learning framework that analyzes facial images and animates them into sounds. By coordinating the movements of the lips and mouth with the sound. It also simulates facial expressions. head movement and even invisible body movements as well
However, it still doesn’t work perfectly. Especially small details such as fingers and teeth. If you look closely, you will find that they still move unnaturally.
As for deepfakes technology, there is currently concern that it will be used by fraudsters to deceive people by impersonating family members or friends through social media. Because today everyone posts pictures and clips of themselves on social media. Therefore it is very easy. If scammers were to save images and sound from the clip to make Deepfakes
Microsoft’s research team is aware of the possible breach, saying: “We have no plans to release additional online demos, APIs, products, usage details.” or any offer related Until we are sure that technology will be used responsibly and in accordance with appropriate regulations.”
Origin techspot.com
–U.S. Air Force Successful test of AI controlling fighter jet Fighting human pilots
#Microsoft #VASA #creates #Deepfakes #photos #audio #tracks