A Hybrid Skin Detection Technique for Surveillance System based on Background Subtraction and Skin Color

Human skin detection is a very popular in many applications, particularly in video surveillance. Many researchers have conducted research in this area, by producing a variety of effective methods. In this paper, research carried out by using a simple algorithm, lower cost, and high detection rate; a hybrid method for detecting human skin region that base on background subtraction and skin color. Background subtraction method is used to detect moving object relying the background model that is calculated by the system. There are three basic steps included in this method. Firstly, a background model is established according to the temporal sequence of the frames. Secondly, the moving objects are detected based on the difference between the current frame and the background model. Finally, the background model is updated periodically to adapt the variety of the monitoring scene. Result show that object can be detect with some noise, but that noise will not effecting the skin detection method in the final result. Next method is skin color method that are used to detecting human skin region in the object detected. Object detected color will be converting from RGB color space to YIQ color space. Individual I channel from YIQ show the high contras between skin and non skin. That I channel are selected to apply the threshold value to obtain the skin mask in binary form. Binary image show that black pixel region is skin region, while white pixel is non skin region. Finally skin mask will use to apply to the original image to define the skin region. The results of the research showing human skin were detected well and noise can be automatic remove after skin color detecting. However, for this paper, the research is carried out only on the skin tan, with indoor environment under white lighting using fixed camera.