PELTE: Privacy Estimation of Images from Tags

Abdurrahman Can Kurtan, Pinar Yolum

Abstract: Image sharing is a service offered by many online social networks. In order to preserve privacy of images, users need to think through and set the privacy settings for each image that they upload. This is difficult for two main reasons: First, research shows that many times users do not know their own privacy preferences, but only become aware of them over time. Second, even when users know their privacy preferences, specifying these policies is cumbersome and requires too much effort, interfering with the quick sharing behavior expected on an social network. Accordingly, this paper proposes an agent-based approach, PELTE, that predicts the privacy setting of images using their content tags. Each user agent makes use of the privacy settings that its user have set for previous images to predict the privacy setting for a new uploaded one automatically. When in doubt, the agent analyzes the sharing behavior of other trusted agents to make a recommendation to its user about what is private. Contrary to existing approaches that assume a centralized online social network, our approach is distributed and thus each agent can only view the privacy settings of the images that it has shared or those that have been shared with it.