Proposal of a Sight-seeing Map with Visual Classification and Presentation of Photos on the Web
暂无分享,去创建一个
In recent years, to search travel information, many kinds of services available on the Web are used widely. For instance, on a photo-sharing website Panoramio, users can browse photos arranged on a map. However, simply arranging a large number of photos may not match a user’s purpose, especially when travelling. Therefore, we have previously proposed a “Sight-seeing Map”, on which the landscapes to which many people pay attention can be intuitively understood by users who plan to travel. On the Sight-seeing Map, photos taken in an arbitrary specified region are classified by scenery categories such as mountain and city, and then a scenery category to which a majority of photos are classified is selected to represent the region, and displayed on the map as an icon. Also actual photos in the region are shown on the Sight-seeing Map. The usefulness of the Sight-seeing Map was confirmed through a questionnaire, for people who wish to plan a travel.
[1] Matthijs C. Dorst. Distinctive Image Features from Scale-Invariant Keypoints , 2011 .
[2] Gabriela Csurka,et al. Visual categorization with bags of keypoints , 2002, eccv 2004.
[3] Brian Everitt,et al. Cluster analysis , 1974 .
[4] Krista A. Ehinger,et al. SUN database: Large-scale scene recognition from abbey to zoo , 2010, 2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition.