The ever present fear of acts of terrorism and crime against persons has caused a dramatic growth in sales of video surveillance systems; for example, in the UK alone, over 4 million closed-circuit TV (CCTV) cameras are installed.
This growth, though, is not enough to guarantee a higher security for citizens, as only a small fraction of the recorded video footage can be examined by trained specialists.

To solve this problem, the Belfast CSIT (Centre for Secure Information Technologies) is studying the possibility of enabling CCTV cameras mounted on buses and trains to instantly recognize suspicious behavior (such as shouts, abusive gestures or language) and dangerous situations, and immediately notify the control center through an Internet-based connection.

At this point, the operator can decide whether he can intervene personally, inform the police or use the train’s PA system to inform passengers that police is on the way, thus restoring calm and avoid worse consequences.

At the current stage, the project is working on improving analysis techniques, aimed at ensuring the system’s automatic intervention only in situations of real danger. In order to achieve this, an artificial intelligence system is being studied, which can draw the correct conclusions from the analysis of behaviors and body language of the people involved, so that non-dangerous information can be filtered out, and the control center be alerted when an intervention is really needed.

The system will evaluate situations according to time of the day, location, moves and gestures of people, assigning a priority level; situations at higher risk and priority will immediately be displayed on the control center screens.
The system should hopefully be in function within 5 years.