¹School of Computer Science and Technology, Shandong Economic University, Ji’nan 250014, China
²Shandong Provincial Key Laboratory of Digital Media Technology, Ji’nan 250014, China

---

Abstract—In this paper, we study on how to automatically mine landmarks from large-scale social images with rich metadata. Firstly, location name is submitted to social image community, and then related social images with rich metadata are obtained. Afterwards, these social images are clustered according to different kinds of metadata of images, and candidate landmarks are mined from the images clustering results. Next, noisy landmarks are pruned from candidate landmarks by computing geographical entropy and time entropy. Experiments conducted on Flickr photos demonstrate the effectiveness of the proposed approach and our approach can also provide useful information for tourists to make tourist plans.

Index Terms—Landmarks, Social Images, Co-clustering, Metadata Similarity.

[PDF]

Cite: Zheng Liu, Hua Yan, and Huijian Han, "Mining Large-Scale Social Images with Rich Metadata and Its Application," Journal of Software vol. 7, no. 4, pp. 749-756, 2012.