Collective evolutionary indexing of multimedia objects

Abstract

The evolution of multimedia technology and the Internet boost the multimedia sharing and searching activities among social networks. The requirements of semantic multimedia retrieval goes far beyond those provided by the text-based search engines technology. Here, we present an collaborative approach that enables the semantic search of the multimedia objects by the collective discovery and meaningful indexing of their semantic concepts. Through the successive use of our model, semantic concepts can be discovered and incorporated by analyzing the users' search queries, relevance feedback and selection patterns. Eventually, through the growth and evolution of the index hierarchy, the semantic index can be dynamically constructed, validated, and naturally built-up towards the expectation of the social network. © 2009 Springer Berlin Heidelberg.