Artificial intelligence has once again demonstrated its prowess in the field of art analysis, this time revealing a hidden detail in Raphael’s painting, the Madonna della Rosa. A team of researchers from the UK and US employed a custom AI algorithm, trained on verified works of the Italian master, to analyze the painting in unprecedented detail. Their findings suggest that while most of the painting can be attributed to Raphael, the face of St Joseph likely wasn’t painted by him.
The research utilized a combination of ResNet50—a deep learning architecture developed by Microsoft—and a Support Vector Machine to scrutinize the artwork. This method, which has been successful in identifying Raphael’s work with 98 percent accuracy in the past, was adapted to examine individual components of the painting. It confirmed that the faces of the Madonna, the Child, and St John were indeed by Raphael, but St Joseph’s face did not match the master’s style.
Such insights are crucial, as they can inform discussions on artwork provenance and authenticity. This technology, while not replacing traditional methods like pigment analysis or provenance studies, acts as a valuable tool in the art historian’s arsenal. The research, detailed in the journal Heritage Science, highlights the evolving role of AI in cultural heritage, aiding experts rather than usurping their roles.
To read more about how AI is uncovering the secrets of classic paintings, visit Science Alert.