The art critic in the machine tells forgeries from the real thing

RESEARCH HIGHLIGHT 22 November 2021

Paintings of flowers honed a neural network’s ability to link a work to a specific artist. Credit: F. Ji et al./Herit. Sci. (CC BY 4.0)

Artificial neural networks can tell one artist’s work from another’s using the 3D topography of the brushstrokes, researchers have found1. The technique could help to identify forgeries, or to distinguish the hands of various members of an old master’s workshop who have contributed to one painting. doi: https://doi.org/10.1038/d41586-021-03447-w
https://www.nature.com/articles/d41586-021-03447-w