SLIF: Structured Literature Image Finder

SLIF is the Structured Literature Image Finder` (previously, *Subcellular Location Image Finder), which was project from the Murphy Lab, which I was a part of. It searches through scientific journals, looking for images, which it categorises, makes available, and (partially) analyses.

SLIF Pipeline

The figure above (from Coelho et al., 2010) is a diagram of the SLIF pipeline: figures are extracted from papers, broken up into panels & captions, which are then analysed to build navigavtion models.

In this project, I was responsible for the computer vision aspects of the processing pipeline. I also handled much of the integration effort of the several components and coordinated the preparation of the multi-author publications that resulted from the effort. This project was a finalist in the Elsevier Grand Challenge (4 teams out of 70 were chosen for the final). I represented our team in both the semi-final and the final of this competition.

The SLIF page on the Murphy Lab website many more details, including some historical information and presentation materials.

References

These are the most recent references, on which I was involved. The first paper focus most on the image analysis side of the pipeline, the second of the overall service.

  • Structured Literature Image Finder: Extracting Information from Text and Images in Biomedical Literature by Luis Pedro Coelho, Amr Ahmed, Andrew Arnold, Joshua Kangas, Abdul-Saboor Sheikh, Eric P. Xing, William W. Cohen, and Robert F. Murphy, in Lecture Notes in Bioinformatics vol. 6004 [PDF version] This also introduced parameter-free threshold adjacency statistics, a variation on threshold adjacency statistics without any parameters (details on pftas).
  • Structured Literature Image Finder: Parsing Text and Figures in Biomedical Literature by Amr Ahmed, Andrew Arnold, Luis Pedro Coelho, Joshua Kangas, Abdul-Saboor Sheikh, Eric Xing, William Cohen and Robert F. Murphy in Web Semantics: Science, Services and Agents on the World Wide Web [online version]

In BibTex format:

@article{j.websem255,
    title = {Structured literature image finder: Parsing text and figures in biomedical literature},
    author = {Amr Ahmed and Luis Pedro Coelho and Andrew Arnold and Joshua Kangas and Abdul-Saboor Sheikh and Eric Xing and William Cohen and Robert F. Murphy},
    journal = {Web Semantics: Science, Services and Agents on the World Wide Web},
    volume = {8},
    number = {2-3},
    year = {2012},
    publisher = {Elsevier},
    keywords = {Image search; Image indexing; Topic modelling},
    issn = {1570-8268},
    url = {http://www.websemanticsjournal.org/index.php/ps/article/view/255}
}

@article{Coelho2010,
    title = {Structured Literature Image Finder: Extracting Information from Text and Images in Biomedical Literature.},
    author = {Coelho, Luis Pedro and Ahmed, Amr and Arnold, Andrew and Kangas, Joshua and Sheikh, Abdul-Saboor and Xing, Eric P. and Cohen, William W. and Murphy, Robert F.},
    doi = {10.1007/978-3-642-13131-8\_4},
    issn = {0302-9743},
    journal = {Lecture notes in computer science},
    month = jan,
    pages = {23--32},
    pmid = {20628552},
    url = {http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=2901994\&tool=pmcentrez\&rendertype=abstract},
    volume = {6004},
    year = {2010}
}

Article filed in categories: Work