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

The following article describes two approaches to determining the geo-coordinates of the recording place of Flickr videos based on textual metadata. The systems are tested on the MediaEval 2010 Placing Task evaluation data, which consists of 5091 unfiltered test videos and metadata records. The first system is a data-driven approach that uses a heuristics based on the spatial variance of tags. The second one extends this heuristics by using semantic technologies, such as extended Word net and a geographical gazetteer. The performance peaks at being able to classify 14% of the videos to within an accuracy of 10m. The article present the two algorithms, evaluates their accuracy and discusses the advantages and disadvantages of using Semantic technologies for this task.