Software Design Of Face Recognition Terminal Based On Android Platform

As graduated into modern society,artificial intelligence has developed rapidly,and the performance of smart device terminals has been greatly improved.At the same time,various identification technologies used for identity verification have gradually entered people’s daily lives,such as password verification,signature verification,credit card verification,and so on.However,the problems of password security,scene limitations,and operational complexity in these traditional identity verifications have always plagued people,and face recognition technology in the same field can effectively solve these problems,making face recognition technology It has received widespread attention.Because traditional face recognition systems are all based on desktop computers,problems such as inconvenience to carry and difficulty in operation may arise.In this paper,the design and research of the face recognition system on the mobile terminal is not only low in cost and convenient to carry,but its application prospects will be more extensive.This paper designs a face recognition system terminal based on the Android platform.The system terminal is mainly composed of three modules:face image registration module,face attribute detection module and face similarity comparison module.First,build the required software and hardware platforms;research image preprocessing technology to effectively reduce the impact of external environmental factors such as color,lighting,accessories,etc.on face recognition;research face detection technology based on Adaboost algorithm,and determine the face based on Haar features Then use the Adaboost algorithm to train a strong classifier,and use the decision tree to combine the strong classifiers into a cascaded classifier for face detection operations to draw a face frame to find people Face area,so as to determine whether there is a face in the picture;use a cross-layer connection module based on deep learning to detect face attributes,and use this to detect face attribute information such as age,gender,and 3D angle of the face;research people Face feature comparison technology,using Gabor wavelet to extract features of the face region obtained according to the face detection operation,and then obtain the similarity between the two by calculating the distance between the feature vectors of different face images,if the similarity is close 1 is recognized as the same face,otherwise it is recognized as different faces;research on live detection technology,including visible light(RGB,Red-Green-Blue)live detection,infrared(IR,Infrared Radiation)live detection and the combination of the two Three modes of binocular live detection together,and then analyze the construction,advantages,disadvantages and corresponding environment of these three modes respectively to maximize the safety and accuracy of recognition;finally,the face recognition system software is designed and implemented according to the built platform and related face recognition technology combined with Open CV library functions.Finally,the face recognition terminal software was tested in the later stage of the project.The system realizes the function of face image attribute detection;realizes the function of face attribute detection based on video frames;realizes the function of comparing pictures and pictures for similarity;realizes the face of video frame and the face in the face library Picture comparison function.On the basis of ensuring the safety of the system,all design functions have reached the expected indicators.The detection speed of the system is relatively fast,and it only takes about 150 ms to complete the face detection.The detection accuracy rate is high,reaching more than98%,which achieves a result that satisfies users.