Bradford, England — In a significant breakthrough for art analysis, researchers have harnessed artificial intelligence to examine a Raphael painting, revealing that a face traditionally attributed to the Italian master may have been created by another artist. The investigation focused on St. Joseph, depicted in the upper left corner of the “Madonna della Rosa,” a work believed to date back to between 1518 and 1520.
The authenticity of “Madonna della Rosa” has long been contested among scholars. Various debates have emerged surrounding the painting’s attribution, and now, with AI technology, findings indicate that at least some of the strokes in the painting could belong to a different creator. A custom algorithm developed by a collaborative team from the UK and the US utilized deep feature analysis to parse through recognized Raphael masterpieces, allowing the AI to identify unique stylistic elements such as brushstrokes, color schemes, and shading techniques.
Hassan Ugail, a mathematician and computer scientist at the University of Bradford, explained that the AI system is capable of analyzing the minutiae of a painting at levels beyond human detection. This technology builds upon the ResNet50 architecture developed by Microsoft, refined with a traditional machine learning technique known as a Support Vector Machine. The model was trained on a vast number of authenticated Raphael works and boasted an impressive accuracy rate of 98% in identifying his paintings.
When testing the “Madonna della Rosa” as a cohesive piece, the results yielded inconclusive results concerning St. Joseph. However, individual testing of the characters revealed that while the Virgin Mary, Child, and St. John bore attributes consistent with Raphael, St. Joseph’s face stood out as likely not being from his brush. The face has previously been perceived as of lesser quality than the surrounding figures, suggesting a possible contribution from another artist, potentially Giulio Romano, a noted pupil of Raphael.
The research underscores a growing trend of utilizing modern technology to uncover hidden truths about historical artworks. The “Madonna della Rosa” has long intrigued art historians, with suspicions regarding its authenticity arising as early as the mid-1800s. This new analysis offers concrete support to those theories, but the team insists that AI will not replace traditional art authentication methods.
“This is about supplementing expertise, not substituting it,” Ugail stressed. The process of verifying artwork authenticity encompasses a range of factors, including provenance, materials, and the painting’s condition. The AI serves as a tool, enhancing art experts’ capacities while maintaining the human element crucial to art history.
The findings were presented in the journal “Heritage Science,” marking a pivotal moment in the intersection of art and technology. As researchers continue to explore the capabilities of AI in art analysis, the future may hold even more revelations about the works of masters like Raphael.