This paper describes how an offline software tool identifies collusion on assessed video clips. The collusion detection processing is independent of the length, format, frame rate, compression and content--of the video clips. The process consists of extracting prominent features from individual frames for each student s video clip and passing these through a rigorous comparison process. Collusion is identified when the software detects similar features when comparing these clips. A summary of results will be presented and the application will then allow--users to view the colluded video via a graphical user interface. This paper also explains how this tool can enhance the learning and teaching when used as a deterrence tool. A software demonstration will--be presented to multimedia students at the start of the academic year. Statistics have shown that students who are aware of the capability of modern collusion detection software will normally be driven to complete--their coursework on their own. The collusion detection tool can also be used in research, where further investigations could be carried out to report on the performance of the detection process.
|Published - 2007