Rapid Face Detection with Multi-pose Knowledge Models and Templates
Ping Shao · Journal of Chinese Computer Systems · 2007
Multi-pose knowledge models are proposed to acquire the candidate face information of size, position, sort of pose and the center coordinates of eyes and mouth from two-value grads image of face organs. According to the sort of pose, the candidate face is affirmed as a true face through template matching with corresponding templates. The face detection, pose estimation and orientation of eyes and mouth are integrated in the algorithm and the detection speed is quite fast. This algorithm is suitable for face detection with multi-pose and multi-face, and since only the gray information of image is used, it can be used for face detection in either gray image or color image.