Tuning the Feature Space for Content-Based Music Retrieval
Proceedings of the 2006 conference on STAIRS 2006: Proceedings of the Third Starting AI Researchers' Symposium
Adaptive content-based music retrieval system
Multimedia Tools and Applications
Hi-index | 0.00 |
A multimedia document retrieval system should be able to handle documents containing elements of multiple media types, both in terms of document representation and query formulation. This paper deals with the problems of extensibility in retrieval of multimedia documents by presenting features of a system capable of handling any media type and formal definitions of two retrieval models supporting this kind of extensibility. The modified extended Boolean model improves query expressiveness from the previous model while retaining the same query computation concepts. The model of similar clusters facilitates the use of information gathered in previous retrieval sessions to improve retrieval results.