Related Experiment Videos
Establishing an international reference image database for research and development in medical image processing
A Horsch1, M Prinz, S Schneider
1Department of Medical Statistics and Epidemiology, Technical University of Munich, Ismaninger Strasse 22, 81675 Munich, Germany. alexander.horsch@imse.med.tu-muenchen.de
Methods of Information in Medicine
|October 9, 2004
Summary
Researchers are developing a reference image database to improve medical image processing (MIP) comparability. This initiative addresses challenges in data standardization for better research outcomes.
Area of Science:
- Medical Image Processing (MIP)
- Health Informatics
- Database Development
Background:
- Lack of comparability in Medical Image Processing (MIP) evaluation results hinders research and development.
- Variations in image dataset quality, size, and gold standards are primary reasons for this lack of comparability.
Purpose of the Study:
- To establish parts of a Reference Image Database for Medical Image Processing (MIP).
- To address the need for standardized datasets in MIP research.
Main Methods:
- Identifying highly relevant medical problems with significant potential for MIP improvement.
- Providing respective reference datasets.
- Defining provider-user relationships through bilateral cooperation or license agreements.
Main Results:
- An explorative database prototype has been implemented using MySQL on the Web.
- Templates for provider-user agreements have been developed and applied in preliminary collaborations.
Conclusions:
- Initial steps have been taken towards a comprehensive reference image database for MIP.
- Key issues such as funding, motivation, management, gold standard provision, and evaluation guidelines require further resolution.
- Continued efforts are planned due to significant interest from research groups and industry.