【作者】 李丹； 郭小华； 刘正熙；

【Author】 LI Dan, GUO Xiaohua, LIU Zhengxi (College of Computer Science, Sichuan University, Chengdu 610064, China)

【机构】 四川大学计算机学院；

【摘要】 基于DSP构建的自动人脸识别门禁系统具有超高速、高性能、高可靠性的优点,本文用Adaboost统计学习方法进行人脸检测和用非线性SVM分类器进行人脸验证,然后用PCA算法和SVM进行识别,最后根据识别结果,判断用户是否合法,控制门的开关。更多还原

【Abstract】 A design of human face recognition algorithm based on DSP realizes an automatically controlled gates system in real time, which has the advantages of super high processing speed, performance and dependability. On the basis of human face detection and confirmation respectively adopting Adaboost statistical learning method and non- linear SVM classifier, .principal component analysis (PCA) and SVM are used to recognize human faces. According to the result of human face recognition, DSP identifies customers and controls the switch of gates.更多还原