People counting system based on skin color and face detection technology

People counting systems aim at automatically estimating the number of people indoor and outdoor places. They are widely used in commercial environment; determining conversion ratio, advertising and promotional evaluations, and so on, video surveillance and transportation management system. In this thesis, two people counting systems were developed; one is based on face detection and the other is based on skin color and face detection technology. Skin color and face detection technology based people counting system (System 2) was developed by adding skin color detection system to the people counting system based on face detection (System 1). This thesis presents a people counting system model, based on skin color detection and face detection technology, counting the number of potential learners who are watching the bulletin board in educational environment. Face detection system of the proposed system uses Viola-Jones method to detect face. Viola-Jones method can detect human faces very quickly and achieve high detection accuracy by using Haar classifier technique. This system was developed using OpenCV library. Widely used 2 color spaces; HSV and YCbCr color spaces were used for skin color detection in the thesis. Experiments showed that the successful results were obtained from HSV color space instead of YCbCr color space, in accordance with the given parameters. As a result of outcomes of this study, it has shown that people counting systems can also be extended in educational applications in addition to such people counting system studies on video surveillance, calculating the number of readers of billboard, etc?KeywordsPeople counting system, People counting, Face detection, Skin color detection, Bulletin board, Computer vision.