Attractive social image extraction based on users' social behaviors

Emerging social networks cause an explosive growth in the available multimedia contents on web. Exploiting these contents and their attached meta-data to improve users' experience in various fields become a major research interest in past years. In this study, a new method to extract attractive images from these networks has been proposed. Also, a case study on Facebook and Flickr has been done to approve the reliability of our approach. By using the meta-data socially attractive images have been collected and used in a questionnaire with 91 participants. According to our results, it has been concluded that extracting social images with bookmark counts more than the average of the bookmarks of the corresponding concept, results in gathering attractive social images.