Scholars studying the painting “Madonna of the Roses” by Raphael have long been debating the authenticity of the artwork. Some thought that the masterpiece, featured at Madrid’s Prado Museum, was indeed the work of Raphael, while other critics were more skeptical and noted some inconsistencies in the painting.
Now, a research paper published in recent months in the journal Heritage Scienceis arguing that one of the subjects in the Madonna of the Roses may in fact not have been the work of the Italian renaissance artist.
As explained by Artnews, the authors of the study, led by Professor Hussan Ugail, a visual computer expert at the University of Bradford in the UK, used an AI algorithm to study the paint strokes of “Madonna of the Roses.”
The algorithm, which was trained on 49 artworks by Raphael and has an accuracy of 98%, found that most of the painting matches the style of Raphael. However, it raised questions on the figure of Joseph, depicted on the left side of the canvas.
“The AI program analysis of our work has demonstrated conclusively that whereas the three figures of the Madonna, Jesus, and St. John the Baptist are unequivocally by Raphael,” Howell Edwards, Professor Emeritus of Molecular Spectroscopy at the University of Bradford told the Guardian, “that of St Joseph is not, and has been painted by someone else.”
“Madonna of the Roses,” an oil on canvas painting completed between 1518 and 1520, depicts the Virgin Mary holding Child Christ with the young St. John the Baptist on the left and St. Joseph in the background. For years, art experts attributed the work to Raphael, with the likely contribution of his pupils Giulio Romano and Giovan Francesco Penni.
One of the most accepted theories posited that the lower part of the painting, the part with the rose, was added at a later stage by one of Rapahel’s aides, but the new AI-powered research confirmed that that portion of the canvas is likely the work of Raphael.
What the algorithm flagged as inconsistent with the Renaissance master’s style was a part of the painting previously attributed to Raphael.
“Using deep feature analysis, we used pictures of authenticated Raphael paintings to train the computer to recognize his style to a very detailed degree, from the brushstrokes, the color palette, the shading and every aspect of the work,” Professor Hassan Uguail explained in a press release. “The computer sees far more deeply than the human eye; to microscopic level. ”
While acknowledging that we will still need the “human eye” of art experts for artwork attributions, Uguail thinks that machine learning algorithms will become a useful tool to decode authors’ painting styles.
“This is not a case of AI taking people’s jobs. The process of authenticating a work of art involves looking at many aspects, from its provenance, pigments, condition of the work and so on,” Uguail said. “However, this sort of software can be used as one tool to assist in the process.”