Sign In

Communications of the ACM

ACM TechNews

Journals Adopt AI to Spot Duplicated Images in Manuscripts

View as: Print Mobile App Share:
A woman reviews the software's results.

Artificial intelligence software that spots duplicated images in research papers can work faster and on a larger scale than manual checkers, but still needs editorial oversight.

Credit: Laurence Dutton/Getty

The American Association for Cancer Research (AACR) and at least three other publishers are vetting manuscripts for duplicate images using artificial intelligence.

AACR's Daniel Evanko said the publisher selected Israel-based Proofig's software to screen submissions for duplicates before publication.

The software extracts and compares images from papers in pairs to find common characteristics, including partial duplications.

Proofig's Dror Kolodkin-Gal said the software can correct for issues like compression artifacts.

Editors still must decide what to do when the software cites images; if datasets are deliberately shown twice, with explanations, then repeated images might be suitable, while some duplications might not be fraud, but caused by copy-and-paste errors.

From Nature
View Full Article


Abstracts Copyright © 2022 SmithBucklin, Washington, DC, USA


No entries found