Design And Implementation Of Face Recognition System In Video Surveillance

Face recognition technology is an important part of the field of artificial intelligence.It has always been a research hotspot in the field of computer vision and pattern recognition.It has important theoretical value and broad application prospects.With the outbreak of deep learning in 2006,Indispensable component of deep learning Convolutional Neural Network(CNN),its application in face recognition is dotted and has made many breakthroughs.In recent years,face recognition technology has been widely used in many fields such as identity verification,secure payment,security system,life and entertainment.This paper mainly studies convolutional neural network algorithm and face recognition algorithm,and implements related algorithms.It integrates various modules of face recognition.The main work includes the following aspects: First,face recognition and face detection The traditional correlation algorithms are studied and studied,and the advantages and disadvantages of several algorithms are analyzed.Secondly,the basic theory and network structure of convolutional neural networks are studied,relevant algorithms are studied in depth,and several of them are applied to face recognition.Convolutional neural network models are analyzed and compared in depth.Thirdly,this paper studies GoogLeNet's face recognition method and builds a face recognition system which includes face detection module,key point location module,pre-processing module and feature extraction and comparison module.Each module is deeply studied.And implementation.Among them,this paper has added a CNN-based key positioning module to the traditional face recognition system to improve the recognition accuracy.This article measures the accuracy of the system and tests the function implementation of each module.Based on the FDDB(Face Detection Data Set and Benchmark)database and the LFW(Labeled Faces in the Wild Home)database,the accuracy of the face detection module,the feature extraction and comparison module,and the entire face recognition system is tested,and the accuracy of the face recognition system is measured.Compare other algorithms.The accuracy of the recognition method in this paper is better than the accuracy of human eye recognition.The accuracy rate of this method is 97.83%,which verifies the performance of this system in face recognition.