Design And Implementation Of Face Recognition System Based On Real-Time Video Data

With the rapid upgrade of computer hardware and the innovation and optimization of relevant theories in the field of face recognition,face recognition has changed from the previous theory to reality,and the recognition accuracy and recall rate have also been greatly improved.Compared with other biometric technologies,face recognition has obvious features,non-contact,easy to forge,and high security.Therefore,it has been widely used in various systems such as surveillance systems,financial systems,forensic systems,and payment systems.It is very convenient for people's work and life.This paper designed a face recognition system based on real-time video data.The system integrates face image acquisition,information management,face detection and recognition functions.It has high accuracy and recall rate in both face detection and recognition.The main work of this paper is as follows:(1)The overall scheme of the system was designed,and the data collection function,data management function,face detection and recognition function were realized.(2)For the Haar cascade classifier,only one face can be detected at a time,and there is a problem of low detection accuracy and low recall rate when the face is deflected and is blurred,a face detection scheme combining OpenCV and target detection algorithm SSD is designed.This scheme can detect multiple faces at the same time,and has higher accuracy and recall rate than the four Haar cascade classifiers.(3)For the traditional LBPH face recognition algorithm,when the face information is extracted,the microscopic information and macro information of the face cannot be taken into account,and the key area information of the face is not complete enough.A multi-scale and key block is proposed,the weighted MBK-LBPH face recognition algorithm has a certain improvement in accuracy and recall rate compared with the traditional LBPH face recognition algorithm.