No Image x 0.00 + POST No Image

AI Unmasks a Hidden Hand in Raphael’s Madonna della Rosa: St Joseph’s Face May Not Be Raphael’s

SHARE
0

An AI trained to read brushstrokes has identified a startling detail in the Madonna della Rosa. The top-left face of St Joseph may not be painted by Raphael at all. The rest of the painting still reads as Raphael, but Joseph’s features appear to be by another hand. The finding fuels a long-running debate about the painting’s true maker.

AI Unmasks a Hidden Hand in Raphael’s Madonna della Rosa: St Joseph’s Face May Not Be Raphael’s

How the AI Reads Raphael’s Style

Researchers built a custom analysis using deep feature analysis. They trained the computer on pictures of authenticated Raphael paintings to learn his brushwork, color palette, shading, and every detail. The computer sees far more deeply than the human eye, down to microscopic features, researchers say. The project emphasizes that the AI’s perception goes beyond ordinary sight.

How the AI Reads Raphael’s Style

The Technology Behind the Finding

To build the system, the team adapted a Microsoft architecture called ResNet50 and paired it with a traditional machine-learning technique called a Support Vector Machine. This approach has previously shown about 98 percent accuracy in identifying Raphael paintings. Usually trained on whole pictures, the researchers also asked the algorithm to examine individual faces.

The Technology Behind the Finding

What the Findings Say About St Joseph's Face

When the della Rosa is analyzed as a whole, results are inconclusive. But when researchers tested individual parts, the rest of the painting matched Raphael, while Joseph’s face did not. The study notes that Giulio Romano, one of Raphael’s pupils, might have painted the fourth face, but this is far from certain.

What the Findings Say About St Joseph's Face

Implications for Art Authentication

Painted between 1518 and 1520, the Madonna della Rosa has attracted suspicions since the mid-1800s that Raphael did not paint all of it. The researchers emphasize that AI will assist human experts, not replace them. Authenticating a work requires provenance, pigments, condition, and other factors, and AI can be a useful tool among many. The research was published in Heritage Science, with an earlier version of this article appearing in December 2023.

Implications for Art Authentication