Call him Vincent van Bot: a team at MIT has created an artificial intelligence that can faithfully reproduce works of art. Dubbed RePaint, the system uses a combination of 3D printing and deep learning to authentically recreate complex paintings.
It could be used to remake famous artworks for the home, or even to produce accurate copies to protect originals from suffering wear-and-tear when hung in museums.
“The value of fine art has rapidly increased in recent years, so there’s an increased tendency for it to be locked up in warehouses away from the public eye,” said RePaint mechanical engineer Mike Foshey. “We’re building the technology to reverse this trend, and to create inexpensive and accurate reproductions that can be enjoyed by all.”
The team used RePaint to reproduce a number of oil paintings created especially for the project by an artist. They developed a technique called ‘color-contoning’, which involves using a 3D printer to lay down translucent inks in thin layers, much like the wafers in a Kit-Kat. They combined this with an established technique called ‘half-toning’, where an image is created by lots of little ink dots, rather than continuous tones. This more fully captures the nuances of the colours in the original works.
At this time, the reproductions are only about the size of a business card, due to the time-costly nature of printing and the limited number of inks available to them.
In the future, the team expects that more advanced, commercial 3D printers could help them to make larger paintings.