AUTOMATIC AND ADAPTIVE FACE TRACKING USING COLOR-BASED PROBABILISTIC NETWORK

Face tracking has potential applications in a wide range of commercial products such as face recognition system. Commonly used face tracking algorithms can extract faces from images accurately, but they often take a long time to finish the detection process or require too much human intervention in system initialization. Recently, there is an increasing demand of robust and fast face tracking algorithm in applications like video surveillance system. This paper aims at proposing a color modeling and testing scheme so that the robustness and applicability of the tracking system can be increased. Experimental results show that the proposed system can detect and track faces from images reliably and quickly.