What is it about?

Wherein your humble protagonists face down technical challenges, stodgy skeptics, and the fear of math to develop an artificial intelligence (AI) system that finds elusive features in different types of images. Images of paintings are difficult to analyze -- for purposes of authentication and attribution -- because there are typically so few of them relative to the voracious appetite of AI systems for training data. Moreover, AI systems need high resolution to analyze subtle image features, such as the artist's brushstroke, but can handle only small images. Similar limitations plague efforts to analyze medical images, e.g., to find telltale signs of cancer in large X-ray images and pathology slides. By dividing a high-resolution image into small tiles and confine the computer's analysis to the ones that pack the biggest visual punch, we achieve high accuracy with a small set of training images. We analyze Salvator Mundi, the world's most expensive painting, and question the received wisdom surrounding a famous painting thought, until recently, to be the work of Rembrandt.

Featured Image

Read the Original

This page is a summary of: State of the Art: This Convolutional Neural Network can Tell you Whether a Painting is a Fake, IEEE Spectrum, September 2021, Institute of Electrical & Electronics Engineers (IEEE),
DOI: 10.1109/mspec.2021.9531029.
You can read the full text:

Read

Contributors

The following have contributed to this page