Face detection in complex background based on gaussian models and neural networks

Abstract

Face detection is a valuable topic in the fields of pattern recognition and computer vision. A Gaussian model and neural network based human face detection method is proposed for color images in complex background. First, Gaussian models are used for the fast extraction of skin-color and lip-color areas. Then, the face candidates are rotated in image plane to be upright. Finally, a neural network determines whether a candidate area is a human face or not. The experimental results illustrate the effectiveness of the method