Design And Implementation Of Face Recognition System Based On Convolutional Neural Network

With the rapid improvement of the capacity of large-scale data storage and the computing power of hardware devices,as well as the continuous innovation of algorithm in the field of neural network,deep learning has rapidly entered a new explosive period from the period of rapid development.The various researches related to deep learning have developed vigorously,and not only have made remarkable progress in theory,but also have promoted the intelligent development of various industries through the organic combination with the existing Internet technology,which has exerted profound influence on the society.The face recognition technology based on deep learning has become one of the most promising research directions in the future due to its ease of use in various identity authentication scenarios.Compared with fingerprint identification method,face recognition acquisition is nonmandatory and does not require direct contact with the device.Compared with the password authentication method,face recognition is very simple and easy to use,which can bring a good human-computer interaction experience.However,human faces are subject to variability and similarity,and face recognition technology is usually more susceptible to interference from environmental factors.This makes it difficult to implement a highly available face recognition system with high security.In this paper,a high availability face recognition system in multi-scene is designed.The work of this paper is mainly composed of the following parts.The first part is the research and improvement of the face detection module.This part mainly analyzes and studies the face detection algorithm,and then improves the convolutional neural network used to extract facial features in the algorithm,so that the face detection module has higher accuracy and faster operation speed.The second part is to realize the core module of face recognition.This part mainly combines and instantiates the image preprocessing module,face detection module,face alignment module,and face feature extraction and comparison module to realize the core module of face detection,feature extraction,face recognition and other functions.The third part is to realize the human face recognition system,this part is mainly to achieve the core module integrated into the whole face recognition system,and realize the communication based on TCP protocol module,log module,control module and other modules,as well as the introduction of database system,the third party send module HTTP request and response,so as to realize can provide for human face recognition based on TCP and HTTP protocol service system.The whole system can provide reliable face recognition service in multiple scenarios,so that the identity authentication function based on face recognition can be realized in PC,mobile phone and other multiple terminals by calling the service.