Enhancing privacy and security within social networks
Abstract
The introduction of online social networking has allowed people across the world to share information with each other. It has risen to be one of the most popular forms of internet usage where almost everyone has created one or multiple types of social network accounts. Along with its great benefits has come many concerns though, mainly that the overabundance of information has created a plethora of privacy issues for not only the users of online social networks, but also for the company hosting the specific social network. Due to the wide range of privacy concerns associated with online social networks, it is incredibly difficult to tackle all the current possible concerns. In this thesis, we propose works to tackle two privacy issues associated with online social networks. These two privacy issues are: the friend search engine, and image content. Firstly we will introduce a new sub-graph approach to the friend search engine that removes the ability of attackers to gain more information of your friend list than your privacy settings allow. Secondly we will introduce a new privacy setting that allows users to define locations they do not wish their face to be seen in images. If an image is posted with their face in such a location, they will be privatized through facial replacement so that they are unrecognizable. The overall efficiency of these works will be tested so that their enhanced privacy does not cause usability issues if they are adopted by a social network site. These works allow users to remain more private while using social media and also help users to remain confident that their privacy is kept safe. These improvements not only help strengthen the privacy of users, but also help social network sites retain users that are more wary of privacy breaches online.
Degree
Ph. D.