Open-i

A service of the United States National Library of Medicine that provides search and retrieval of biomedical images from the full text and figure captions of articles in the open access subset of the biomedical literature database. The platform indexes over four million images including photographs, medical illustrations, radiological images, microscopy images, charts, graphs, and tables extracted from published research articles, making it one of the largest publicly accessible collections of biomedical visual content. Users can search for images using text queries that match figure captions and article text, or they can perform visual similarity searches by uploading a reference image and retrieving visually similar images from the indexed collection. The text-based search supports filtering by image type such as chart, photograph, MRI, CT, ultrasound, microscopy, or illustration, as well as by publication date, journal, and MeSH subject headings. Each image result links back to the source article in the full-text repository, providing context for how the image was used in the original research. The platform supports natural language processing of figure captions to extract semantic concepts and entities, enabling more intelligent image retrieval than simple keyword matching. The visual similarity search uses computer vision algorithms to compare image features, helping researchers find comparable cases or reference images for their own work. Researchers in radiology, pathology, dermatology, and other visually oriented medical specialties use the system to find diagnostic reference images.