Learning landmarks by exploiting social media

  • Authors:

  • Chia-Kai Liang;Yu-Ting Hsieh;Tien-Jung Chuang;Yin Wang;Ming-Fang Weng;Yung-Yu Chuang

  • Affiliations:

  • National Taiwan University;National Taiwan University;National Taiwan University;National Taiwan University;National Taiwan University;National Taiwan University

  • Venue:
  • MMM'10 Proceedings of the 16th international conference on Advances in Multimedia Modeling
  • Year:
  • 2010

Quantified Score

Hi-index 0.00

Visualization

Abstract

This paper introduces methods for automatic annotation of landmark photographs via learning textual tags and visual features of landmarks from landmark photographs that are appropriately location-tagged from social media. By analyzing spatial distributions of text tags from Flickr's geotagged photos, we identify thousands of tags that likely refer to landmarks. Further verification by utilizing Wikipedia articles filters out non-landmark tags. Association analysis is used to find the containment relationship between landmark tags and other geographic names, thus forming a geographic hierarchy. Photographs relevant to each landmark tag were retrieved from Flickr and distinctive visual features were extracted from them. The results form ontology for landmarks, including their names, equivalent names, geographic hierarchy, and visual features. We also propose an efficient indexing method for content-based landmark search. The resultant ontology could be used in tag suggestion and content-relevant re-ranking.