Video and image identification: tools for supporting forensic investigations

The world's law enforcement agencies are increasingly tracking down producers and distributors of offensive multimedia content such as child abuse material. The challenge lies in the vast amount of easily accessible images and videos on the Internet, as well as the large capacities of today's hard drives. Inspecting each and every file is infeasible in practice, given police investigators' limited resources and time. Additionally, such inspection work is repetitive and slow, and in the case of deeply offensive images even psychologically harmful. In the presentation we will first review the standard process and challenges that police investigators currently go through when investigating the contents of common storage devices and the challenges connected with this work. Second, we will present the currently available technical toolkits that can assist police investigators and save significant time during this tedious research process. We will focus especially on tools capable of extracting and classifying the content from large collections of multimedia files as this is the most time consuming task for investigators. Finally we will look at the problem in a larger context, as not only police departments can benefit from the development of tools helping to automatically identify and/or classify multimedia content. Several other organizations are also involved in the fight against the distribution of offensive multimedia content on the internet, e.g. INHOPE (Internet Hotline Providers in Europe) and NGOs fighting for childrens' rights, as well as associations of internet service providers such as EUROISPA.