AUTOMATIC SECURITY SYSTEM THROUGH BODY ACTIONS AND FACIAL EXPRESSIONS VIA REAL TIME IMAGE PROCESSING
Advisor:
Dr. Kashif Javed
Abstract:
Terrorism has become an International problem from the past 6 years. Almost every country is
affected by it in some way or another. With the increased crime rate the need of security has
become more imminent than ever before. Conventional security systems of camera networks
based on CCTV technology or even IP cameras are not capable to provide response of an
activity of interest in real-time systems. A software level solution was required to assist the
security personnel in their task of narrowing down the suspicious person from different screens
in front of him. To enhance the precision and focus point of the security personnel we have
developed our system.
Our proposed solution consists of two major parts. The facial expressions and body gestures. In
the facial expression part, we will be utilizing the facial gestures of the person to identify if he is
exhibiting suspicious behavior in the facial expressions e.g. either he is sad or confused or
frightened. Our system also tracks the body posture if that person is suspicious i.e. one of the
gestures that has been identified as suspicious is detected, then this person will be marked and
security personnel will be informed about him and he will check that person carefully so that he
is caught before performing any dangerous activity. In this way the scope of view of the security
personnel has been narrowed down from multiple screen to one screen and then from multiple
people in the scene to one single person.