AdaBoost algorithm has been approved in fast face detection by more people. However, this algorithm often has false detection in case of the complex environment. In order to achieve both fast and accuracy detection, a face detection method with human eye location algorithm as the auxiliary algorithm is proposed. At first, face detection is realized using the AdaBoost algorithm, then the fast recognition by the improvement template match algorithm to the human eyes. In the human eye detection, the examination can be sped up using the pyramid algorithm. The experiment shows that this method can reduce the false detection rate of the AdaBoost algorithm.%在快速人脸检测算法中,基于AdaBoost算法的人脸检测算法得到越来越多的认可.然而,该算法在复杂环境下常有误检的情况.为了既快速又能提高检测的准确率,提出了以人眼定位的算法作为辅助的人脸检测算法.首先应用AdaBoost算法检测人脸的位置,然后再利用改进的模板匹配算法对人眼进行快速识别与定位.在人眼检测方面,利用金字塔算法加快了检测速度.实验表明,该方法可减少AdaBoost算法的误检率.
展开▼