Multi-View Face Detection Based On Skin Color And Adaboost

Face detection is a hot spot in computer vision and has attracted much attention because of its wide applications in many areas such as face recognition and human computer interaction. In this paper, a multi-view face detection method is proposed, which combined skin color information and AdaBoost-based face detection technique together to improve the detection accuracy and speed. The.contributions of this paper includes:1. Large numbers of face samples of different poses and non-face samples have been collected from videos and images and a face sample collection has been set up. Then serveral Adaboost classifiers have been trained on the collection with each detector responsible for one view.2. An efficient and robust method for multi-view face detection is proposed. The possible face regions are extracted by skin color first, and then multi AdaBoost face detector are applied to the possible face regions to get the exact face position. As much area of the image is excluded by the skin-color information, the speed of the algorithm is improved greatly.3. A face detection method is proposed which can detec faces on skin binarized images based on AdaBoost detectors training on skin binarized face images. It can strengthen the difference between different regions in the face and reduced the account of the needed features.