Data Science in Art: Discerning the Painter’s Hand

The Physics, Materials Science and Engineering, and Art History and Art Departments at Case Western Reserve University have joined together in collaboration for the “Data Science in Art: Discerning the Painter’s Hand”. 

The goal of this project is to apply machine learning methods to high-resolution 3D scans of the surfaces of paintings in order to attribute stylistic components of brushwork to artists. Specifically, this team is studying later paintings of El Greco with the aim of distinguishing between his brushstrokes, those of his son Jorge Manuel and other workshop participants, as well as later conservation interventions.  

Ken Singer, the Ambrose Swasey Professor of Physics, Michael Hinczewski, the Warren E. Rupp Associate Professor of Physics, Ina Martin, Adjunct Assistant Professor of Materials Science and Engineering, and Elizabeth Bolman, the Elsie B. Smith Professor in the Liberal Arts and chair of the Department of Art History and Art, recently completed a controlled experiment to use machine learning on a series of painted images of lilies by artists from the Cleveland Institute of Art.

In order to simulate workshop methods and produce sufficient training data, this group of students painted three versions of the same photo of lilies. Twelve of these paintings (4 groups of 3 paintings) were selected. Then, the team compared the paintings by scanning the surfaces of the paintings with an Optical Profilometer, thus creating high-quality, quantitative scans of the surface height distributions. The machine was trained on two of the three paintings by each student. We used 200 x 200 pixel virtual patches that were unknown to the machine to test attribution. 

The team was able to attribute 98% of the patches, a significant result given that the patch size was a fraction of a complete brush stroke. The methods made it possible to attribute brushwork confidently even when the virtual patches were as small as the width of a few bristles. 

This project confirmed that these methods can be successful on recently made paintings. Next, our team will apply these methods for attribution on El Greco’s paintings, in collaboration with Factum Arte, the premier organization working at the intersection of art and technology.