figure search logo             Biomedical Figure Search is a search engine which finds figures based on the caption.

 

Biomedical Figure Search is a search engine which finds figures based on the caption. As of May 2013, we have indexed about 790,000 biomedical articles that were published as open access. In addition to a search engine, the system also provides a summary for each figure. This summary is generated automatically by selecting the most relevant sentences from the full-text.

 

Enter your search query and hit "Search". You'll get a set of results displayed, with the article title, abstract and figure caption. You can also view the automatically generated summary of any figure below its caption.

The summary is made up of 3 sentences. All sentences of the article are classified as either background, method result or conclusion. The summary displays one background sentence, one results sentence and one conclusion sentence.

 You can also click on the article's title to view other figures in the article.

  •  A video demo of FigureSearch is also available.
  •  Please provide your opinion by taking this short survey.

Why is Biomedical FigureSearch limited to ~800,000 articles?

Our FigureSearch indexes the captions from the open access articles. Due to copyright restrictions and the non-availability of many articles, we cannot display information from a lot of articles. PubMed Central, a free library of biomedical articles, makes available a set of open access articles. Currently, the size of this set is about 790,000 articles.  As more articles are published in the open access domain, we will be able to index them and improve the quality of our results.

How is the summary generated?

We have developed a technology to automatically detect sentences from full-text of the article that are most relevant to the figure. We present 3 such sentences, one Background sentence, one Results sentence and one Conclusion sentence. We believe that these sentences together provide a brief overview of the figure. These sentences are selected on the basis of words that are present in the caption and the sentence.

Browser extensions

Figure summarization is available as a browser extension now for Firefox, Chrome and Safari. On installing this extension, when you look at a figure in PubMed Central from an open-access article, a summary will be inserted under the caption of the figure (see figure below).

Without Extension                                                                             With Extension

withoutextension         withextension

Extension Download Links:

Supporting this project

The most important requirement for the success of this project is availability of articles published without copyright restrictions, in other words, open access articles. The best way to support this project would be to publish your articles though one of the open access journals. A list of open access journals can be found at - http://www.doaj.org/.

 

Journal papers

  • Lee M, Wang WQ, and Yu H. 2006. Exploring supervised and unsupervised methods to detect topics in biomedical text. BMC Bioinformatics 7(1):140. link
  • Yu H and Lee M. 2006. Accessing bioscience images from abstract sentences. Bioinformatics 22(14): e547-e556. link
  • Yu H, Agarwal S, Johnston M, and Cohen A. 2009. Are figure legends sufficient? Evaluating the contribution of associated text to biomedical figure comprehension. BMC Journal of Biomedical Discovery and Collaboration.4:1. link
  • Agarwal S and Yu H. 2009. Automatically classifying sentences in full-text biomedical articles into Introduction, Methods, Results, and Discussion. Bioinformatics 25(23):3174-80. link
  • Kim D and Yu H. 2010. Automatic figure classification in bioscience literature. Journal of Biomedical Informatics (submitted).
  • Agarwal S and Yu H. 2010. Biomedical negation scope detection with conditional random fields. Journal of American Medical Informatics Association. link
  • Yu H, Liu FF, Ramesh BP. 2010. Figure ranking, novel user interface and natural language processing approaches for automatic figure ranking. PLoS ONE. link
  • Kim D and Yu H. 2010. Figure text detection in bioscience literature. PLoS ONE (accepted).

Conference papers

  • Yu H and Lee M. 2006. BioEx: A novel user-interface that accesses images from abstract sentences. The Annual Conference of the North American Chapter of the Association for Computational Linguistics (NAACL-HLT), New York. link (pdf)
  • Yu H. 2006. Towards answering biological questions with experimental evidence: Automatically identifying text that summarize image content in full-text articles. American Medical Informatics Association (AMIA) Fall Symposium, Washington, DC. link
  • Rafkind B, Lee M, Chang SF, and Yu H. 2006. Exploring text and image features to classify images in bioscience literature. The Annual Conference of the North American Chapter of the Association for Computational Linguistics (NAACL-HLT) Linking Natural Language Processing and Biology: Towards Deeper Biological Literature Analysis (BioNLP), New York. link
  • Zweigenbaum P, Demner-Fushman D, Yu H, Cohen KB. 2007. New frontiers in biomedical text mining. Pacific Symposium on Biocomputing (PSB), Big Island of Hawaii. link (pdf)
  • Cohen KB, Yu H, Bourne PE, and Hirschman L. 2008. Translating biology: text mining tools that work. Pacific Symposium on Biocomputing (PSB), Big Island of Hawaii. link (pdf)
  • Agarwal S and Yu H. 2009. Automatically classifying sentences in full-text biomedical articles into Introduction, Methods, Results, and Discussion. American Medical Informatics Association (AMIA) Summit on Translational Bioinformatics, San Francisco, California. link (pdf slides)
  • Agarwal S and Yu H. 2009. Automatically generating structured text summaries for figures in biomedical literature. American Medical Informatics Association (AMIA) Fall Symposium, San Francisco, California. link
  • Kim D and Yu H. 2009. Hierarchical image classification in the bioscience literature. American Medical Informatics Association (AMIA) Fall Symposium, San Francisco, California. link

Posters

  • Agarwal S and Yu H. 2009. Adding Automatically Generated Summaries to Biomedical Figures to Improve Literature Search. Presented at AMIA 2009 Annual Symposium, San Francisco, CA
  • Agarwal S and Yu H. 2010. Improvements in the Figure Search and Summarization System. To be presented at AMIA 2010 Annual Symposium, Washington, DC

Watch a video demo |  | Lucene query syntax

Project supported by NIH grant 1R01GM095476

Contact us: This email address is being protected from spambots. You need JavaScript enabled to view it.

Search powered by Lucene. Developed at UWM and hosted at AWS

This page might NOT be displayed correctly in Internet Explorer Bottom of Form