Research And Design Of Anti-spoofing Face Recognition Attendance System

With the development of pattern recognition and machine learning,face recognition technology has made great progress.Compared with other biometric technologies,face recognition has the incomparable advantages and a wider application field,has been became a research hotspot in machine vision field.The main task of this paper is taking research on attendance system that based on face recognition technology,and proposed improvement plan to allow quick and accurate spoof detection which from photos and video fraudulent attacks.First of all,this paper introduces the common methods of face detection and its performance comparison,using AdaBoost algorithm to complete the real-time face detection and localization;Through the human eyes positioning technology,Pupil coordinates was precisely determined;Through positioning coordinates of the eyes,achieved the geometric transformation,scale normalization and histogram equalization preprocessing of the face images,and complete face region image acquisition.Secondly,this paper analyzes the different methods of face liveness detection and their applicable conditions,Usingthe Lambertian model,adopted scheme based on single image without adding any auxiliary equipment to conduct face liveness detection;the LBP texture feature and Fourier spectrum feature detection scheme were introduced detailed,this paper proposed to combining LBP texture feature and Fourier spectrum feature as a new feature extraction scheme;taking Support Vector Machine as classifier and NUAA Imposter Database as testing database,Primary experiments on a large photo imposter database show that the proposed method gives preferable detection performance compared to others.Then,the common face recognition algorithm is analyzed,and its applicability and use environment are compared.This paper described principal component analysis(PCA),two dimensional principal component analysis(2DPCA)and linear discriminant analysis(LDA)in detail;proposed to combining 2DPCA with LDA algorithm as a new recognition algorithm;took the ORL Database as algorithm testing database to test algorithm.The result show that the fused plan have a higher recognition rate and faster recognition time.Finally,designed and constructed the system.This paper analyzed the requirement of the Attendance System,programmed module,database and interface,tested system.The testing result show that system is stable and reliable,and has met the design requirement.