Semantic integration and knowledge exchange for diagnostic medical image databases

Adrian S. Barb,C. Shyu,Yash Sethi
DOI: https://doi.org/10.1109/BIBE.2004.1317340
2004-05-19
Abstract:Information technology offers great opportunities to radiologists to utilize their expertise in decision support and training. Collaborative approaches in these areas enable physicians to access relevant cases diagnosed by experts from other health care groups. Unfortunately, there is little agreement on a single model of semantic representation and information exchange. In this context, semantic interoperability among heterogeneous groups plays an important role in a collaborative setting. In this paper, we propose a model for semantics integration and knowledge exchange in collaborative environments that feature heterogeneous semantics integration. It provides a computational and visual mechanism to associate synonymous semantics of visual abnormalities related to lung pathologies. We also offer a solution for system level communication that improves the retrieval precision using peer domain expertise. From our experiments we obtained a high degree of matching between different semantics that describe the same visual pattern of lung pathology. Also, our experiments show that, using the knowledge exchange mechanism, the default system setting adjusts well over time to increase the retrieval precision for new users.
What problem does this paper attempt to address?