Every chart, figure and microscope image in a scientific paper tells a story but sometimes, that story is fake. Copy-move forgery, where parts of an image are duplicated to fabricate results, is one of the most common and damaging forms of scientific misconduct.

In this competition, you’ll build models to detect and segment these forgeries in biomedical images. Using a dataset of confirmed forgeries from over 2,000 retracted papers, your models will identify duplicated regions at the pixel level and help flag tampered images before they spread.

Total Prizes:

• $55,000

Entry Deadline:

• January 8, 2026