Data-Driven vs. Semantic-Technology-Driven Tag-Based Video Location Estimation

TitleData-Driven vs. Semantic-Technology-Driven Tag-Based Video Location Estimation
Publication TypeConference Paper
Year of Publication2011
AuthorsChoi, J., & Friedland G.
Page(s)243-246
Other Numbers3219
Abstract

The following article describes two approaches todetermining the geo-coordinates of the recording place of Flickrvideos based on textual metadata. The systems are tested on theMediaEval 2010 Placing Task evaluation data, which consists of5091 unfiltered test videos and metadata records. The first systemis a data-driven approach that uses a heuristics based on thespatial variance of tags. The second one extends this heuristicsby using semantic technologies, such as extended Wordnet and ageographical gazetteer. The performance peaks at being able to

Acknowledgment

This work was partially supported by funding provided through the National Geospatial-Intelligence Agency University Research Initiatives program (NGA NURI #HM11582-10-1-0008). Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors or originators and do not necessarily reflect the views of NGA.

URLhttp://www.icsi.berkeley.edu/pubs/speech/datadriven11.pdf
Bibliographic Notes

Proceedings of the Fifth IEEE International Conference on Semantic Computing (ICSC 2011), pp. 243-246, Palo Alto, California

Abbreviated Authors

J. Choi and G. Friedland

ICSI Research Group

Audio and Multimedia

ICSI Publication Type

Article in conference proceedings