Real-time human skin color detection algorithm using skin color map

Human computer interface (HCI) system requires a skin color detection algorithm in it's first step of designing. Skin color detection algorithm faces two major problems one is variable Human skin tone and second one is complex background. Human skin tone variation can be easily seen from person to person within same ethnic group or different ethnic groups. And the background used in real time is usually complex with variable lightening conditions. Hence a good skin color detection algorithm is required to work well under these conditions. In this paper a skin color detection algorithm using skin color map is proposed which gives promising results in complex background. Color space used here is YCbCr color space and skin color model used is explicitly defined regions. This model has certain advantages over parametric and non parametric models in terms of implementation and memory usage and color space chosen is good for video processing applications.