SINAI at Placing Task of MediaEval 2010
José M. Perea‐Ortega, Miguel Ángel García Cumbreras, Luís Alfonso Ureña López, Manuel Garcı́a-Vega · 2010
In this paper we present a basic approach for assigning geographical coordinates to videos uploaded to Flickr 1 . Our approach combines geographical entities recognition based on the annotations provided for each video and information retrieval on textual tags of Flickr images provided for development purposes. However, we have not used the low-level visual features extracted from keyframes and training images. The poor results obtained show that our geographical entity recognizer must be improved to solve the spatial ambiguity present in most of the user annotations of each video. In addition, some problems have been identied during the