Research On Face Detection And Recognition Algorithm Based On Computer Vision

With the continuous development of technology and the continuous innovation of science and technology,the application of network technology is becoming more and more extensive.While the network technology is convenient for our daily life,the problems that have been exposed are also emerging in an endless stream.Among them,the most prominent problem is the issue of information security.At present,face detection and recognition technology is developing rapidly.It has been widely applied in e-commerce,security monitoring,e-government,and has achieved excellent performance in intelligent security,intelligent monitoring,electronic office,human-machine interaction and so on.After a large number of data collection and demand analysis,this paper mainly outlines the development process and development status of face detection and recognition technology.From the two aspects of face detection technology and face recognition technology,the related mature algorithms are introduced and analyzed.In the face detection technology,this paper introduces the convolution neural network and the Opencv method in detail.Convolution neural network is an important branch of machine learning.The Opencv method is a set of more mature algorithms,which is more stable in face detection.In the face recognition technology,this article from three aspects,classified and summarized the existing open-source methods and techniques,including: Seetaface method,YouTu method and Facenet method,introduces the method and principle of the three kinds of open source methods involved in each module and on this basis,their advantages and disadvantages.The emphasis of this paper is divided into two parts,one part is the DeepID face detection and recognition based on the model of training,training by using the DeepID method of face detection and recognition model,and through the method of adding joint Bayesian model,in order to improve the recognition rate of the improved DeepID method.The second part is by comparing the various methods of face detection and recognition proposed in the above,pointing out its advantages and disadvantages in the practical application of face detection and recognition,and puts forward further improvement measures.And through to include face detection and recognition algorithm,haze angle offset,facial facial occlusion and facial exaggerated facial expressions of four kinds of demand under the condition of comparative tests,various situations are integrated in the same system,selecting the best choice in the case of a best algorithm as the processing algorithm under the condition of the same system in which can handle many problems at the same time,the maximum extent to improve the detection rate and recognition rate,and reducing the false detection rate.