A Realistic Framework for Counter-terrorism in Multimedia

The internet has become an easy platform for video broadcasting by providing an inexpensive but weak publishing barrier to everyone and eventually attracting the huge audience. Terrorists make use of videos as an efficient medium for spreading their message; showing violence to attract sympathies, terrify viewers and promoting radicalization; videos act as an important weapon towards their mission. Existing internet security solutions and blocking methods rely on textual data and lack analysis of visual content. In this paper, we propose a novel framework that can realistically automate the screening of such videos using visual content analysis and consequently alert the authorities. After examining existing discrete definitions of violence, we classify the broad spectrum of violence in videos into four streams which serve as the core functions of the proposed system. The system is capable of evaluating the semantic context as well as the extent of violence in the videos. Being fully automatic and comprehensive, the system works efficiently and more effectively in comparison to any existing systems proving itself as a powerful combating tool in controlling the means and effects of this mode of terrorism.