Skip to Main Content
Face detection is a preliminary step for a wide range of applications such as face recognition, video-surveillance and so on. The object of this work is to improve the correct rate of face detection. Skin-color model is established first to extract the possible face region, then the BP(Back Propagation) neural network model is used to simulate the output of the possible human face region and Bayesian decision theory is used to classify the face or non-face pattern. Experiments show that for color face images using skin-color model is effective and fast; the use of BP neural network simulation to remove the dummy distinguish is an effective way; the use of Bayesian decision theory to analyze the overall co-ordination, correct rate of face detection has been further improved.