Leonardo Da Vinci’s infamous Mona Lisa has always been a subject to various technological experiments. Mona Lisa is greatly known for her enigmatic smile, on which art scholars have spent years to decipher the real meaning behind. With the birth of AI based technologies like Deepfake, Mona Lisa has been doing way more than just smile. Deepfake is an AI based technology that produces realistic animation of human heads that is considered as a tedious task because of the complexities involved. “Human heads are geometrically complex and highly dynamic; 3D models of heads have tens of millions of parameters,” the study authors wrote. Machine learning experts at have made a system which can make animations and small videos of motion by using just one photo or painting.
In a recent video shared on Youtube, three video clips show examples of Mona Lisa moving her head and lips and it looked extremely realistic. The animation was created by a type of AI called a convolutional neural network that processes information like the human brain does while viewing and analyzing images. The algorithm was trained by the researchers to interpret facial features and its general shapes, this information was later applied to still images. Facial movements from datasets of three humans, producing three different animations, were used for the making of the Mona Lisa video. The three clips portrayed Mona Lisa but with the behavioral pattern of the three personalities.
According to the study, the human vision system can identify even “minor mistakes” in animations of 3D modeled human heads. In order to make the animation as realistic as possible, the technology has to be fed with multiple angles of the desired subject with references of actual human faces in action. The AI would then be trained to map the movement of the source’s features. Videos that used around 32 images achieved what was called as “perfect realism” which portrayed an extremely realistic animation of the subject.