摘要
人脸检测作为自动人脸识别系统的第一个环节具有非常重要的作用,为了解决目前大部分人脸检测方法存在的分类器训练困难和检测计算量大等问题,提出了一种人脸检测的混合方法。该方法由两级分类器组成,第一级为粗分类器主要过滤大部分非人脸区域,第二级为核心分类器,在由第一级粗分类的基础上利用非线性SVM算法进行人脸检测。在CMU数据库上的实验结果表明,该方法具有较高的人脸检测率,检测速度得到大幅提高。
Face detection is the first step in the automatic system of face recognition and plays an important role in the system. However, there are some problems in the previous methods of face detection, such as the difficulty in classifier＇s training and calculation. An improved hybrid face detection method is proposed. The new method includes two different classifiers. The first one is called cursory classifier, which filtrates most of the non-face windows. The second one is a core classifier, which uses non-linear SVM method to detect the face based on the result from the first classifier. Experimental results from CMU face image database show that the efficiency of this method is high and the detection speed is improved greatly.
出处
《华东船舶工业学院学报》
北大核心
2005年第4期68-72,共5页
Journal of East China Shipbuilding Institute(Natural Science Edition)
基金
中国科学院沈阳自动化研究所机器人学重点实验室基金(RL200108)
江苏省青年科技基金(BQ2000315)
江苏省高校自然科学研究计划项目(01KJB520002)
江苏省自然科学基金(BK2002001)。
关键词
人脸检测
支持向量机
统计学习理论
模式识别
face detection
support vector machine
statistical learning theory
pattern recognition
