Prosemantic Image Retrieval
In this technical demonstration we present a content-based image retrieval system based on the ‘query by example’ paradigm. The system effectiveness will be proved for both category and target search on two standard image databases, even without a “good” initial example and ancillary information, such as device metadata, text annotations, etc. These results are obtained by incorporating in the system our recently proposed prosemantic features coupled with a relevance feedback mechanism, and by maximizing novelty and diversity in the result sets.
KeywordsImage Retrieval Relevance Feedback Relevant Image Text Annotation Ancillary Information
- 2.Santini, S., Cusano, C., Schettini, R.: Diversity and novelty in multimedia search. In: ACM Multimedia 2012 (submitted, 2012)Google Scholar