CEA LIST's Participation at MediaEval 2012 Placing Task

Multimedia content geotagging is potentially useful in a wide variety of application. Although an increasing number of existing devices include geotagging options, a wide majority of online content is still not geotagged and methods for efficient automatic geotagging are needed. Here we describe CEA LIST’s participation to the MediaEval 2012 Placing Task [3]. We submitted runs that exploit either textual or visual information to place videos on the map and we will briefly introduce the methods used. The main innovation with respect to state of the art methods was to create user geotagging models and to rely on them if useful textual annotation are missing. Exploiting user models proves efficient since geotagging accuracy is significantly improved compared to the use of generic language models and/or gazetteers.