基于CNN的人脸反欺诈的框架研究与实现

Research and Implementation of CNN-based Face Anti-fraud Framework

随着5G的发展和信息技术的推广,人脸识别和反欺诈越来越受到重视,但市面上现有的人脸识别系统在技术上存在采集数据不完整、识别风险管理存在漏洞等问题,同时人为的欺诈行为也越来越常见。为防止人脸识别系统将假人识别为真实人脸用户,本文提出基于卷积神经网络的人脸反欺诈算法框架,最终实现自动化的人脸反欺诈识别技术。实验结果表明,该方法的准确率达到73.23%。

With the development of 5G and the promotion of information technology,face recognition and anti-fraud are getting more and more attention.However,the existing face recognition systems on the market have technical problems such as incomplete data collection and loopholes in recognition risk management.At the same time,human fraud is becoming more and more common.In order to prevent the face recognition system from recognizing fake people as real face users,this paper proposes a face anti-fraud algorithm framework based on convolutional neural network,and finally realizes automatic face anti-fraud recognition technology The experimental results show that the accuracy of our proposed method has reached 73.23%.