Discerning the painter’s hand: machine learning on surface topography
F. Ji, M. S. McMaster, Samuel Schwab, Gundeep Singh, L. N. Smith, Sayanta Adhikari, Marcio O'Dwyer, Farah Sayed, A. Ingrisano, D. C. Yoder, E. S. Bolman, Ina T. Martin, M. Hinczewski, Kenneth D. Singer
Abstract
Abstract Attribution of paintings is a critical problem in art history. This study extends machine learning analysis to surface topography of painted works. A controlled study of positive attribution was designed with paintings produced by a class of art students. The paintings were scanned using a chromatic confocal optical profilometer to produce surface height data. The surface data were divided into virtual patches and used to train an ensemble of convolutional neural networks (CNNs) for attribution. Over a range of square patch sizes from 0.5 to 60 mm, the resulting attribution was found to be 60–96% accurate, and, when comparing regions of different color, was nearly twice as accurate as CNNs using color images of the paintings. Remarkably, short length scales, even as small as a bristle diameter, were the key to reliably distinguishing among artists. These results show promise for real-world attribution, particularly in the case of workshop practice.