Abnormal behavior detection for early warning of terrorist attack

Many terrorist attacks are accomplished by bringing explosive devices hidden in ordinary-looking objects to public places. In such case, it is almost impossible to distinguish a terrorist from ordinary people just from the isolated appearance. However, valuable clues might be discovered through analyzing a series of actions of the same person. Abnormal behaviors of object fetching, deposit, or exchange in public places might indicate potential attacks. Based on the widely equipped CCTV surveillance systems at the entrance of many public places, this paper proposes an algorithm to detect such abnormal behaviors for early warning of terrorist attack.