PERSONALIZED GEO-TAG RECOMMENDATION FOR COMMUNITY CONTRIBUTED IMAGES

Prasanna Wadekar, Santosh Shinde

Abstract: Tagging is popularized by many social sharing websites, which allows us to add the description to object. Using tags users can organize their data so that it will be helpful for searching and browsing. Geotagging of a photo is the process in which a photo is marked with the geographical identification of the place it was taken. Geotagging can benefit users, to discover an Extensive Variation of exact Location related information. In personalized tag recommendation, tags that are relevant to the users query are retrieved based upon the users interest. The introduction of the Hypergraph learning is to find joint relevance between the visual and textual domains. Given a photo with Geolocation and without tags, System uses nearest neighbour search to obtain some user- predilected tags and geo-location predilected tags individually. It discovers the semantically and visually related images, and explores the idea of annotation-by-search to recommend tags for the untagged photo. In conclusion, the tags are recommended to the use

Keywords: Hypergraph Construction, Hypergraph learning, Personalization, Geo-tags, Preference learning

DOI: https://doi.org/10.15623/ijret.2015.0405045