On Building a Self-organizing Search System for Multimedia Retrieval

This publication doesn't include Faculty of Arts. It includes Faculty of Informatics. Official publication website can be found on muni.cz.

Authors

SEDMIDUBSKÝ Jan DOHNAL Vlastislav ZEZULA Pavel

Type Article in Proceedings
Conference International Workshop on Multimedia and Semantic Technologies (MUST 2010)
MU Faculty or unit

Faculty of Informatics

Citation
Doi http://dx.doi.org/10.1109/FUTURETECH.2010.5482652
Field Informatics
Keywords self-organizing system; bootstrap techniques; similarity search; multimedia retrieval
Description In this paper, we build a self-organizing search system that operates in an unstructured P2P network and allows users to search for multimedia data by their content. In order to efficiently route queries to relevant peers, we define and evaluate several techniques for joining new peers to the existing network. These techniques create new relationships between peers --- on the basis of answers returned to queries --- so that a new peer will be able to efficiently forward queries and other peers will be immediately informed about its data. In addition, we demonstrate resilience of the system to sudden peer failures by studying system performance and quality of returned answers after a large number of peers is disconnected. The experiments, evaluated on a synthetic and real-life multimedia dataset, confirm that proposed techniques are suitable for dynamic environments.
Related projects: