Leonardo Da Vinci probably never expected his painting of Mona Lisa to become one of the most recognisable and revered pieces in the entire world. At The Louvre in Paris, tourists from around the world come to view the historic work of art. The story behind the model in the portrait still reminds somewhat shrouded in mystery, with not much on the woman in question being recorded. But what if we could see what she could've looked like in reality? Well, it's now possible.
Découvrir notre dernier podcast
More authentic than the real thing
The developers at Samsung’s AI Center chose the famous Italian master's painting to present their progress in the development of new artificial intelligence. Based on a mixture of facial recognition technology and machine learning, the aim is to video scan a human face in order to allow the computer to understand the make of the face and its movements. Based on this, artificial intelligence is then able to simulate movements and expressions and apply them to a still image.
A “fake news” factory?
This is by no means a new process. In 2016, the "Face2Face" project impressed users by showing the face of former US President George W. Bush manipulated in real-time by a researcher. Other examples have since been released on the internet, including two particularly recent and realistic ones created using the faces of Barack Obama and Donald Trump.
This practice dubbed "Deep Fake", created controversy following its use on adult websites where the faces of celebrities were added to the bodies of actresses in pornographic videos. A questionable use of images which is equally as worrying in terms of integrity, as it is in affecting the quality of news, in a time when fake news is spreading more and more easily across social networks.