Sign In

Communications of the ACM

ACM TechNews

Different Strokes: Using AI to Tell Art Apart

View as: Print Mobile App Share:
An illustration of how the system differentiated the work of four different painters.

The researchers scanned the students’ paintings to produce three-dimensional surface height data for each work.

Credit: The Daily - Case Western Reserve University

Researchers at Case Western Reserve University have developed an artificial intelligence algorithm that can distinguish between the brush strokes of different artists.

The tool potentially could improve the identification of forgeries of famous works, among other things.

The new technique reads and learns from a painting's three-dimensional (3D) topography.

The researchers generated 3D surface height data for each painting, divided them into virtual patches, and trained a convolutional neural network to determine which painter made each stroke.

The neural network was able to identify the correct artist 95% of the time from brush strokes about the diameter of a single bristle.

From The Daily - Case Western Reserve University
View Full Article


Abstracts Copyright © 2021 SmithBucklin, Washington, DC, USA


No entries found