Visual summarization of the social image collection using image attractiveness learned from social behaviors

  • Jeong, Jin woo; 
  • Hong, Hyun ki; 
  • Heu, Jee uk; 
  • Qasim, Iqbal; 
  • Lee, Dongho
Citations

SCOPUS

5

초록

How to effectively summarize a large-scale image collection is still an important and open problem. In this paper, we propose a novel method to effectively generate a summary of the social image collection using image attractiveness learned from the social behaviors conducted in Flickr. To this end, we exploit the note information of Flickr images. The notes of Flickr images are user generated bounding boxes with text annotations assigned on the interesting image regions. Using the visual features extracted from the images that have notes, we have generated the attractiveness models for various concepts. Finally, the attractiveness models are exploited to make a summary of the social image collection. Through various user studies on the image collections from Flickr groups, we show the feasibility of our method and discuss further directions. © 2012 IEEE.

키워드

Flickr note; image attractiveness; image interestingness; image summary; search result clustering; Flickr note; image attractiveness; image summary; Interestingness; Search results; Exhibitions
제목
Visual summarization of the social image collection using image attractiveness learned from social behaviors
저자
Jeong, Jin woo; Hong, Hyun ki; Heu, Jee uk; Qasim, Iqbal; Lee, Dongho
DOI
10.1109/ICME.2012.196
발행일
2012-07
유형
Conference Paper
저널명
Proceedings - IEEE International Conference on Multimedia and Expo
페이지
538 ~ 543