Can social tagged images aid concept-based video search?

13 years 2 months ago

Download staff.science.uva.nl

This paper seeks to unravel whether commonly available social tagged images can be exploited as a training resource for concept-based video search. Since social tags are known to be ambiguous, overly personalized, and often error prone, we place special emphasis on the role of disambiguation. We present a systematic experimental study that evaluates concept detectors based on social tagged images, and their disambiguated versions, in three application scenarios: withindomain, cross-domain, and together with an interacting user. The results indicate that social tagged images can aid conceptbased video search indeed, especially after disambiguation and when used in an interactive video retrieval setting. These results open-up interesting avenues for future research.

Arjan T. Setz, Cees G. M. Snoek

Real-time Traffic