摘要
基于生物特征识别的信息安全终端管理系统利用具有活体检测功能指纹识别技术、基于深度卷积神经网络的人脸检测识别身份认证技术以及目标检测跟踪技术，实时地监控信息终端区域目标运动以及入侵物体情况，对不安全状态进行预警，确保信息安全。实验表明，该系统能检测到多角度多姿态复杂光照人脸，人脸识别在LFW（Labeled Faces in the Wild）数据库上获得99．2％的准确率，并能在特定区域内进行有效跟踪预警。
An information terminal security management system based on 5iorAetrics recognition that integrating technology of fingerprint, technology of detection and recognition based on deep learning convolutional neural networks, and automatic tracking-detecting algorithm, is used to monitor and warn changes to fields in real time. Experimental results demonstrate that the system can effective and accurately detect the intrusion objects adapt to the variations of illumination and pose, and the accuracy of our algorithm achieves accuracy rate of 99. 2% on the well-known and challenging Labeled Faces in the Wild（LFW） benchmark.
作者
杨培德
黄仁裕
尤俊生
YANG Pei-de HUANG Ren-yu YOU Jun-sheng(Xiarnen Meiya Pico Information Co. , Ltd,Xiamen 361008, China)
出处
《计算机科学》
CSCD
北大核心
2016年第B12期225-227,共3页
Computer Science
关键词
信息安全
人脸识别
深度卷积神经网络
入侵检测
Information security,Face recognition,Deep convolutional neural networks,Intrusion detection
