Relevance Computation for Keyword Searches over Structured Multimedia Metadata


Multimedia search engines rely heavily on content analysis algorithms and knowledge management techniques to adequately populate metadata databases. However once this metadata is available, the search engine must work on its own to extract the greatest possible amount of information from the user’s query and map it to the available metadata structure. This paper presents a relevance computation mechanism devised for text-based search engines for multimedia data, in which there is a mismatch between the type of user input (keyword searches) and the metadata model (a structured description of a decomposition of the media), built in a system designed for maximum user simplicity, both at the annotation side and at the end-user side.

Extracted Key Phrases

6 Figures and Tables

Cite this paper

@inproceedings{Villegas2005RelevanceCF, title={Relevance Computation for Keyword Searches over Structured Multimedia Metadata}, author={Paulo Villegas and M{\'o}nica D{\'i}ez and Daniel Garcia}, year={2005} }