Real-Time Face Pose Estimation in Video Sequence

An effective and fast face pose estimation method in video sequence has been proposed based on estimating the relative location of mouth in a face. Firstly, Faces are detected in the first several frames of video sequence. Then the detected face region is being tracked by the CAMShift (Continuously Adaptive Mean Shift) algorithm and an ellipse model is used to locate human face regions and angles accurately. By evaluating the relative location of mouth in a face and the direction angle of ellipse model, human face posture is estimated. Result shows the method is effective and can meet the real-time requirement for estimating face pose in video sequence.