At the University of California, Riverside, three scholars have begun a research project testing – for the first time – the use of facial recognition software to identify unknown subjects of portrait and sculpture art.
The project is called FACES (Faces, Art and Computerized Evaluation Systems). Project director and UC Riverside art history professor Conrad Rudolph hopes to use this cutting-edge facial recognition technology to help answer questions of political and social history surrounding unknown portraiture.
“A walk through almost any major museum will show a large number of these portraits from before the 19th century – many of them great works of art – have lost the identities of their subjects through the fortunes of time,” Rudolph said.
Testing will begin by comparing the death or life mask of a known individual to an identified sculptural portrait of the same individual. Using the results of this testing phase, the researchers must work through a number of issues that art mediums pose.
“With portraiture in sculpture, painting and drawing, not only do all the problems that apply to human subjects apply [lighting, angle, age, etc.], but these works of art have their own additional challenges,” Rudolph said. “Most notably, portrait art does not provide what might be called a photographic likeness but rather one that goes through a process of visual interpretation on the part of the artist.”
Furthermore, the type of paint or technique can drastically change the appearance of an individual within a portrait as well. According to UC Davis doctoral student in computer science Divya Banesh, who has looked into similar research ideas, some type of pre-processing procedure would have to be employed in order to “fix” a portrait individual’s likeness, in order to make the portrait recognizable to the software.
“[For example] many paintings that used the Impressionistic or Neo-Impressionistic techniques did not use expanses of color but short strokes or dots of paint,” Banesh said. “The human eye can stitch this together to create a coherent image, but a facial recognition software might not be able to do the same. It might be like a low-resolution pixelated image that needs to be improved in quality.”
If this initial phase of testing from a 3D lifemask to 3D imaging of the portrait is encouraging, Rudolph and his researchers hope to broaden their methods to address the issues posed by 2D images by starting with a comparison of portraits of the same subject by the same artist, then by different artists.
“We would systematically broaden testing from highly controlled paradigms to less controlled,” Rudolph said. “An important goal of FACES is to test identified portraits against unidentified portraits.”
If successful, the impact of this research on the art world would be unprecedented. Jeffrey Ruda, UC Davis art history professor specializing in the Italian Renaissance, looks ahead to wider research opened up by this project.
“By identifying differences from actual human anatomy, the software could help to define what kinds of interpretation show up in various cultures or in the work of individual artists,” Ruda said.
The study not only opens doors to enriching the understanding and learning of art history and its social and political impacts, but also demonstrates a way art can partner with science in an interdisciplinary way.
“This project will bring a new scientific objectivity to a traditionally highly subjective aspect of art history while at the same time retaining the human eye as the final arbiter,” Rudolph said.
RACHEL KUBICA can be reached at science@theaggie.org.
