The British company Flawless has released the program TrueSync, which uses machine learning to synchronize dubbing with the facial expressions of actors. This means that the movements of the lips and facial muscles of people on the screen will coincide with what the voice actors say.
Kyiv. Ukraine. Ukraine Gate – May 9, 2021 – Technology
“TrueSync is the world’s first system that uses artificial intelligence to create perfectly synchronized visualizations in different languages. The system is based on a performance-saving mechanism that takes into account all the nuances and emotions of the original material,” the statement said.
Representatives of the company note that this software package preserves all the emotions as in the original. The system replaces the actor with a dipfeek, taking into account his original facial expressions, so as not to lose the acting.
Deepfake, the concatenation of the words “deep learning” and “fake”, a technique of synthesizing the image of a person based on artificial intelligence. It is used to combine and overlay existing images and videos on the original images or videos.
Deepfake can be used to create fake news and malicious scams.
Deepfake videos can be easily found on popular streaming video websites such as YouTube or Vimeo. A popular program is FakeApp, which uses TensorFlow.
Methods for detecting gestures and converting the original video to video with a person similar to the target person were introduced in 2016 and allow real-time calculation of fake facial expressions in existing 2D video.
Read also: Tiktok Will Pay Americans $92 Million for Privacy Breaches
Source: Ukrgate
