Research On Cross-age Face Recognition Based On Generative Adversarial Network

That face recognition is widely used in real life has advantages of not easy to forge,user acceptance,convenient detection and high reliability.However,an important constraint for face recognition technology is cross-age face recognition,because there is a big gap between the actual age of the person to be detected and the time of collecting facial images stored in the database.There are obvious differences in the information features of the face,which leads to the situation of ineffective recognition or misrecognition,especially in the case of the larger age span,the problem of ineffective identification or misidentification is more serious.In view of the above problems,in this paper,methods of the machine learning in face recognition system is studied and the cross-age face recognition experiment based on machine learning is designed.It is found that the common machine learning method can not achieve the purpose of cross-age face recognition.In view of this problem,generative adversarial network in recent years performs well in image processing and machine vision.In this paper,the generative adversarial network technology are deeply studied and a cross-age face recognition method based on generative adversarial network technology is proposed.The method uses a Conditional Adversarial Auto Encoder(CAAE)to generate facial images of different age groups of the person to be detected and uses the generated images to compare the images with the images stored in the database.The cross-age face recognition rate can be improved by reducing the difference of face features with the increase of age.This paper designs a single-sample face verification experiment with different ages and a multi-sample face verification experiment with a specified age.The experimental results show that when the proposed method is used for cross-age recognition,the recognition accuracy of the simulated face images generated by CAAE in a given age range is significantly higher than that of the cross-age face images without CAAE processing.It provides a novel and feasible way of method for the research of the across-ages face recognition.