Trust-aware Media Quality Profiles in Fake Multimedia Detection
Nowadays, it becomes increasingly difficult to find reliable media content on the Web since lots of interested parties provide fake media. Automatic fake detection methods often lack scalability and the ability to infer forgery. In previous work we argued for a community-based fake media detection approach using Web 2.0 methods. In this follow-up paper we argue that collected trust-aware media authenticity ratings should be aggregated into a media quality profile. Furthermore, we refine our previous work by addressing three aspects. First, we present a simple game-based approach to overcome the cold-start problem inherent in Web 2.0 communities. Second, we identify the notion of confidence, trust, distrust and their dynamics as necessary refinements of our existing model. Third, we improved the computation of trust-aware aggregated media authenticity ratings by introducing a trust inference algorithm for yet unknown sources in our XMPP-based collaboration framework.
Proceedings of the International Workshop on Multimedia on the Web, in conjunction with i-Know and i-Semantics 2011, 7–9 September, Messe Congress Graz, Austria, PP. 27-32, IEEE Computer Society, 2011.