Investigation into the Automated Detection of Image Based Cyber Bullying on Social Media Platforms

Access to the Web 2.0 technologies and social networking has become a ubiquitous factor of our lifestyles. Mobile devices, smart phones and other such technologies create a permanent link between the person and their presence on the Internet. Reporting daily activities, feelings, opinions and emotions have become common place, particularly with teenagers and young adults. Circles of friendships can expand rapidly without careful consideration to what is being shared and to whom. Such an environment has become an avenue for cyber bullying, enabling persecutors to breach the safety of the home environment to reach their victims. In order for social networking to be a safe technology in the lives of adolescents and vulnerable adults, measures need to be put in place to monitor and detect potentially threatening online activity. In this paper an overview is given of the possible techniques that could be employed in the detection of negative online interactions. The combination of sentiment analysis with image processing techniques is considered as a suitable platform from which to categorize the textual and visual connotations of the content. This is illustrated as a flow diagram illustrating the key processes envisaged for the detection of potential cyber bullying threats.