This system processs live video stream from Microsoft Kinect Camera. The System monitors live stream and detects the individual on the screen. First of all, it classifies each individual based on their age and gender. This system is concerned with men between 18 to 40-year-old who are most likely to carry attacks or perform other criminal/suspicious activities. It filters out women, children and the elderly because their possibility of committing such incidents is very low and this is not the focus of the project scope. It processes all the young males in the screen.
It then begins emotion recognition on these individuals and detects the emotions and behavior of all these people. The six basic emotions that will be detected are:
• Happy • Sad • Fear • Surprise • Anger • Disgust
The system then filters out all these males based on emotions, males that are happy will not be focused. However, individuals that are sad, angry, frustrated, afraid will be the main target as these people are most likely to commit acts of violence. These individuals will be the main focus, all their activities will be observed, such as what is their posture? How are they walking? and if they continue to perform suspicious activities such as hit someone, or are in an argument with someone or have an odd appearance (dressing) which might be suspected as a bomb jacket, the securities subscribed to that system will be alerted and the location of the target will be pin pointed so that actions can be taken to stop the suspect. Moreover, the suspect’s data will also be saved in our database which would be useful for the security personnel later on.