Distributed Content-Based Multimedia Information Retrieval Over the P2P Network
(I. King)

This project aims to design and implement a retrieval strategy for multimedia content over a distributed environment. Unlike typical centralized database, peer-to-peer (P2P) network allows users to share information as well as computation resources in a distributed fashion. Although this can be done currently through a textual interface, much work still needs to be done for content-based multimedia information retrieval in P2P, i.e., retrieval by image similarity based on color, texture, shape, and relations, timing code, video expressions, etc. Issues such as automatic organization of multimedia content by similarity through clustering, query propagation and retrieval, and performance improvement on implementation in P2P networks are still under investigation. In this project, we plan to examine two main issues: (1) the design a distributed protocol for content-based multimedia information retrieval and (2) implement such protocol on existing P2P networks and conduct performance evaluation of the experimental system. This work is applicable in digital library, distributed information retrieval, and web engineering projects.