说明：  基于olivettifaces数据库的人脸识别。 传统的人脸识别只是图像处理和机器学习的结合，在基本的人脸考勤和门禁等有约束的条件下，能满足性能要求。如果是在安防实时监控、智能视频人脸分析、海量人脸搜索验证等非约束的情况下，传统的人脸识别就很难达到要求了。基于卷积神经网络的人脸识别通过训练集来学习，能在无约束的条件下进行人脸识别。
(Face recognition based on olivettifaces database. Traditional face recognition is only a combination of image processing and machine learning. It can satisfy the performance requirements under the constraints of basic face attendance and entrance guard. It is difficult to meet the requirements of traditional face recognition if it is not constrained by real-time security monitoring, intelligent video face analysis, massive face search and verification. Face recognition based on convolution neural network is studied by training set, which can recognize face in unconstrained condition.)