Research On Methods Of Face Liveness Detection

With the development of information technology,people pay more and more attention to information security,and online identity authentication technology is gradually developed from the traditional way to biometric authentication.While biometric authentication technology brings convenience to our daily life,it also brings out a new security problem-Counterfeit biometric attack.Therefore,biometric living detection technology comes into being.In recent years,face liveness detection technology has made some progress,but its reliability is too poor to effectively promote business.In order to achieve high security and strong generalization in face liveness detection technology,a face detection algorithm based on replay attack is studied in the paper.There are main works of this paper:1.Firstly,several effective techniques for face liveness detection are discussed;A review of the literature on face liveness detection based on replay attacks is given out,and the characteristics and advantages and disadvantages of each methods are analyzed;Five kinds of face liveness detection databases which are widely used in the literature are listed,and the emphases of different detection databases is analyzed;Several evaluation indexes of face liveness detection technology are summarized.2.Secondly,we study the reliability,security and generalization ability of HRM+HOG in face liveness detection algorithm.Firstly,HRM and HOG features are extracted,and then SVM classifier parameters are adjusted according to binary signal detection theory.Under the condition of small ACER fluctuation,the APCER value is reduced to the maximum to improve the safety of face liveness detection system.The experimental results using NUAA and OULU-NPU databases show that the HRM+HOG face liveness detection algorithm not only has higher correct detection rate,stronger generalization ability,but also has better system security.3.Thirdly,in order to achieve high accuracy of liveness detection in face authentication system,a face authentication system based on special application(FASSA)was designed.A face identification is performed first,and then liveness detection is performed on the face image after recognition.The FASSA system's liveness detection module is experimented,in which LBP,HOG,GLCM,HAAR and other methods are used to extract features.The liveness verification is carried out by establishing different SVMs or extreme learning machines(ELM)for different individuals.The experimental results show that the proposed detection method is effective for specific environment security requirements.All the experimental results show that the proposed face liveness detection algorithm has strong generalization ability,and the facial authentication system in specific environment has strong pertinence and security.