A Linguistic Approach to Aligning Representations of Human Anatomy and Radiology

Pinar Wennerberg, Manuel Möller, Sonja Zillner

In: Proceedings of the International Conference on Biomedical Ontologies (ICBO). International Conference on Biomedical Ontologies (ICBO-09) July 20-26 Buffalo NY United States 7/2009.


In the context of medical imaging different domain ontologies are necessary that provide complementary knowledge about anatomy and radiology. This is essential for realizing applications such as medical image search. Consequently, semantic integration of these different but nevertheless related types of medical knowledge from disparate domain ontologies becomes necessary. In our work we interpret semantic integration as aligning a taxonomy on radiology and an ontology on human anatomy to find equivalent concepts that represent their shared view on medical imaging. The resulting alignments describing this common view can then be used to annotate medical images and related textual patient data. Our alignment approach has three main aspects: (a) linguistic-based, (b) corpus-based, and (c) dialogue-based. In this paper, we describe the application of the first aspect on a representation of human anatomy and a representation of radiology and report on the results.


Weitere Links

Deutsches Forschungszentrum für Künstliche Intelligenz
German Research Center for Artificial Intelligence