������̳

VizioMetrics is an image search engine and classifier created by researchers at the University of Washington. In order to improve it, we would like to automatically identify a “central figure” in a scientific article in cases when multiple figures are present. We defined “central figure” as a single visualization that encapsulates key aspects of a paper, a graphical summary that captures the content of the article for readers at a single glance. We surveyed 488,590 researchers in biomedical field and found out that for an overwhelming majority of papers their authors were able to identify a single “central figure.”