Georeferencing Flickr photos using language models at different levels of granularity: an evidence based approach

Olivier Van Laere, Steven Schockaert, Barth Dhoedt

Abstract


The topic of automatically assigning geographic coordinates to Web 2.0 resources based on their tags has recently gained considerable attention. However, the coordinates that are produced by automated techniques are necessarily variable, since not all resources are described by tags that are  sufficiently descriptive. Thus there is a need for adaptive  techniques that assign locations to photos at the right level of granularity, or, in some cases, even refrain from making any  estimations regarding location at all. To this end, we consider  the idea of training language models at different levels of  granularity, and combining the evidence provided by these  language models using Dempster and Shafer’s theory of  evidence. We provide experimental results which clearly confirm  that the increased spatial awareness that is thus  gained allows us to make better informed decisions, and moreover increases the overall accuracy of the individual  language models.

Full Text: PDF
Type of Paper: Research Paper
Keywords: Dempster-Shafer evidence theory, Language models, Georeferencing, Web 2.0, Geographic information retrieval
Show BibTex format: BibTeX