A system that can help CCTV operators identify people carrying concealed guns is being developed at Loughborough University. The MEDUSA (Multi Environment Deployable Universal Software Application) project will help reduce gun crime by providing software that can analyse CCTV images and report whether people in the frame, are in the frame. The system will work in in real time allowing the police to grab anyone toting a gun before a serious crime is committed.
The research team will examine genuine CCTV footage of people carrying concealed firearms, to identify any characteristics which are associated with the behaviour and activity of criminals before they commit a gun associated crime. This could be body stance/shape, movement or eye contact with cameras. Using this information they will then develop a novel machine-learning system for behavioural interpretation. This software will enable CCTV camera systems to automatically scan footage, searching for any behavioural characteristics which indicate that an individual may be carrying a gun and, if necessary, alert CCTV operators.
The researchers will also develop a training programme for CCTV operators to enable them to improve their ability to detect suspicious characters. For this they will meet with experienced operators and identify the key cues they already utilise to detect people carrying concealed weapons, and then make these available for training purposes. A detailed database of CCTV footage of individuals carrying or not carrying concealed firearms will also be generated, which can then be used by local authorities and the police for training CCTV operators in what to look for. Alastair Gale, Head of Loughborough University’s Applied Vision Research Centre is leading the MEDUSA project: “The primary aim of this research is to make the UK a safer place to live by reducing the occurrence of gun crime,” he says, “To do this we need to develop ways of identifying a potential gun crime before an incident occurs – a key instrument in this is CCTV.”
The new software will significantly improve CCTV systems enabling individuals to be apprehended before they commit a crime. As the CCTV infrastructure is already in place it should be relatively easy to implement MEDUSA. Its introduction would also lead to the more efficient use of police time, as well as more socially acceptable police practices as the identification of potential criminals becomes improved. Additionally, the system will yield faster post-event analysis of CCTV footage.