PUBLICATION
Automatic Semantic Image Annotation within the Domain of Software Modeling
Type
Conference Paper
Year
2010
Authors
Antje Boehm-Peters
Event
Published in
Proceedings of the IADIS WWW/INTERNET
Abstract
The information society requires capable means to access the vast set of digital content. Currently, search engines are the predominant approach to find suitable data. Although state-of-the-art search engines deliver decent results for textual content, algorithms to retrieve appropriate bitmap data fall short on providing proper image selections. Therefore, we propose a semantic image annotation solution that automatically classifies raw bitmap data visualizing software diagrams with an ontological formalism. Rather than applying pattern recognition techniques to an existing image pool, bitmap data is enhanced by semantic annotations at image creation time. Thus, the underlying data model can be exploited to derive semantic annotations. In contrast to pattern recognition algorithms, our approach implies a precise ontological classification which is the foundation of accurate bitmap search.
Reference
Heinrich, Matthias; Boehm-Peters, Antje: Automatic Semantic Image Annotation within the Domain of Software Modeling. Proceedings of the IADIS WWW/INTERNET, 2010.
