Related Experiment Videos
Image informatics at a national research center
L Rodney Long1, Sameer K Antani, George R Thoma
1Communications Engineering Branch, US National Library of Medicine, Lister Hill National Center for Biomedical Communications, National Institutes of Health/DHHS, 8600 Rockville Pike, Bldg 38A MS 55, Bethesda, MD 20894, USA. rlong@mail.nih.gov
Summary
Researchers at the National Library of Medicine (NLM) develop computer-assisted methods for extracting information from document and biomedical images. Their work focuses on journal articles and medical image analysis for improved data retrieval.
Area of Science:
- Biomedical Engineering
- Computer Science
- Information Science
Background:
- The Communications Engineering Branch (CEB) at the Lister Hill National Center for Biomedical Communications (LHNCBC), a research division of the National Library of Medicine (NLM), conducts image informatics research.
- Image informatics encompasses both document and biomedical images.
Purpose of the Study:
- To develop and implement computer-assisted methods for information extraction from various image types.
- To advance research in document image analysis and biomedical image processing.
Main Methods:
- Document image analysis focuses on extracting bibliographic data from scanned journal articles.
- Biomedical imaging research involves content-based image retrieval (CBIR).
- Key areas include segmentation, indexing, and classification of medical image collections.
Main Results:
- Prototype systems incorporating computer-assisted information extraction methods have been implemented.
- Research is ongoing in extracting structured data from unstructured journal articles.
- Advancements are being made in analyzing collections of spine and uterine cervix images.
Conclusions:
- Computer-assisted methods are central to the mission of extracting information from document and biomedical images.
- The CEB's research contributes to improved data retrieval and analysis in both domains.
- Future work will likely involve further refinement of CBIR and data extraction techniques.