Face Recognition Systembased On Deep Learning

Face image recognition mainly refers to automatically infer the identity,expression,age,gender and other attribute information by analyzing the face image,video or the collection of pictures and videos.Face image recognition technology is widely used,but the current face image recognition technology is affected by the background,illumination and lens angle,and the recognition effect is not ideal.In this paper,based on deep learning,face image recognition is studied.The main contents are as followsFirstly,several network structures commonly used in deep learning,such as deep belief network,recurrent neural network and convolutional neural network,are introduced,and the basic structure,working principle and calculation process of convolutional neural network,which is the most commonly used in deep learning,are studied in detail,The working principle and calculation process of convolution neural network are studied.Secondly,on the basis of convolutional neural network,this paper studies the face recognition model based on VGG network,studies the related technologies used in face recognition,and explores the improvement of convolutional neural network for face image recognition,which introduces attention mechanism and triple loss function.Finally,on the basis of the above research content,this paper simulates the face image recognition model based on deep learning,builds the model by using python,trains the model through the corresponding data set,and completes the face image recognition test.Compared with Face Net and Deep Face,the recognition effect is good.Then,the Open CV software library is used to build a real-time face recognition system based on the face image recognition model.Finally,the face detection and recognition in the actual scene of the system are carried out,and good results are achieved.