ATINEGAR-KOOSHA: A Genetic-algorithm-based method for feasibility of face recognition in CCTV cameras
Today, by development of technology in CCTV systems, their applications have become more widespread and specialized. In this article at first, CCTV systems are introduced. After that the components of a CCTV camera and the function of them are described. One of the most popular applications of CCTV systems is their ability to recognize the faces of suspects. What is presented in this article is the introduction of a method based on genetic algorithm by which the feasibility of this event can be investigated. The feasibility has been investigated in different resolution conditions, in sensors with different sizes and at the required variable distances from the camera. This feasibility study has been done with the approach of increasing the pixel density assigned to people's faces for better detection as well as increasing the transverse field of view of the camera. Also in this method based on genetic algorithm, which is called "Atinegar-Koosha" method, all practical limitations can be detected, including the limitation in the focal length of the camera lens, the minimum and maximum transverse angle of view of the camera, the minimum required width for face detection and also level of importance coefficient of the required pixel density - to further enhance the ability to recognize people's faces - they are applied quite accurately.