Ontology-based Annotation and Retrieval System for Digital Mammography Images

This paper describes major components and early results of a working in progress project which includes ontology based annotation and retrieval system for Digital Mammography archive. In this study, we first present three layered system architecture aiming to close well-known semantic gap problem. Then we propose an integrated retrieval model which combines textual, visual and high level features together. Then, we present the ontology we developed for digital mammography in OWL, and summarized studies on preparation of evaluation and training set. Additionally, we present an annotation tool which is used for annotation of evaluation set