Using AI to Restore Historical Murals
MGEN Assistant Teaching Professor Zheng Zheng has been working closely with a high school student to develop an AI tool to help restore historical murals.
This article originally appeared on Northeastern Global News. It was published by Noah Lloyd. Main photo: Assistant teaching professor Zheng Zheng is creating an AI tool to help art restorationists. Courtesy photo.
How one high schooler’s ambition led to an AI tool to restore ancient art
Zheng Zheng started painting when he was 6 years old. Now, as an assistant teaching professor in Northeastern University’s College of Engineering, he wants to use artificial intelligence to restore historic murals.
The inspiration came from a teenager. Qianhao Han was in the 10th grade when she emailed Zheng out of the blue, inquiring how she could get involved in research. The Toronto high school student said she had reached out to several professors, to no avail. But Zheng, whose laboratory is located at Northeastern’s Toronto campus, saw potential in the ambitious young person and replied.
It turned out that Han had a passion for both art and science and had earlier participated in robotics competitions and taught herself the basics of AI. But she craved “more professional guidance,” she said.
Zheng invited her to give a presentation on one of her projects that uses AI to detect pedestrians and signal to cars when they’re crossing the street.
Soon, Han was working in the lab, shoulder-to-shoulder with Zheng’s graduate students — like Jincheng Jiang, a master’s student also working on the mural restoration project.
Using AI to build a tool for art restorers is actually “something that I proposed,” Han says.
The program, which they call MurFact AI, or just MurFact, concludes a mural’s material that might be either missing or distorted and then generates its interpretation of the image.
Advanced technology alongside ancient murals
While the technique applies to many kinds of murals, both Han and Zheng were inspired by those of ancient China. “We love history,” Zheng said, smiling.
Zheng explained that restoring murals is far trickier than restoring other art, like a photograph. This is partly due to the sheer scale of some murals, which can stretch over 10 meters, but also because some murals are displayed outside and exposed to damage from humidity, sunlight or human factors, Zheng said.
Read full story at Northeastern Global News