Term-relevance computations and perfect retrieval performance
Information Processing and Management: an International Journal
Optimization of relevance feedback weights
SIGIR '95 Proceedings of the 18th annual international ACM SIGIR conference on Research and development in information retrieval
Bridging the semanitic gap in image retrieval
Distributed multimedia databases
Learning to map between ontologies on the semantic web
Proceedings of the 11th international conference on World Wide Web
Information Retrieval
Modern Information Retrieval
Ontology-Based Photo Annotation
IEEE Intelligent Systems
Object Recognition as Machine Translation: Learning a Lexicon for a Fixed Image Vocabulary
ECCV '02 Proceedings of the 7th European Conference on Computer Vision-Part IV
MindReader: Querying Databases Through Multiple Examples
VLDB '98 Proceedings of the 24rd International Conference on Very Large Data Bases
SMAP '06 Proceedings of the First International Workshop on Semantic Media Adaptation and Personalization
Relevance feedback: a power tool for interactive content-based image retrieval
IEEE Transactions on Circuits and Systems for Video Technology
Hi-index | 0.00 |
The assignment of annotations still forms the basis for conceptual queries in large image/text repositories by facilitating data classification at semantic level. However, the varying users' perception of image contents and the usage of different retrieval aspects make it necessary to develop methods for the unification and integration of different annotation schemes. In this paper we present the IKONA Retrieval and Annotation System with its main focus on the transformation of the subjective annotations assigned by different users into a unified knowledge base. For that purpose, the conducted queries are adjusted to user-dependent preferences by finding correspondences between the used vocabulary and the system's 'core' annotation ontology. The introduced method is evaluated on a large collection of news data including both images and the corresponding textual data. The experiments show that our approach significantly increases the retrieval quality, particularly when users are faced with a data repository whose content is unknown and has not been made completely semantically accessible.