
Bioscience-scale automated detection of figure element reuse
This study presents an algorithm that detects figure element reuse in a large dataset of 760 thousand open access articles and 2 million figures, robust to various transformations. A panel review estimates that 0.6% of all articles contain fraudulent reuses, with inappropriate reuses occurring 43% across articles, 28% within an article, and 29% within a figure.
Read paper(opens in new tab)
