一种在线考试中实时图像监考系统的设计与实现  

Design and Implementation of Real-time Image Monitoring System for Online Exam

作　　者：黄志威[1] 宋鸿陟[1] 司国东[1] 

机构地区：[1]华南农业大学数学与信息学院,广州510642

出　　处：《现代计算机》2015年第13期62-67,共6页Modern Computer

基　　金：国家自然科学基金项目(No.60875045)

摘　　要：针对在线考试中一些常规的作弊行为以及当前视频监控方法存在的问题,提出一种基于图像的在线考试监考方案,设计并实现一个实时图像监考系统。该系统通过Web RTC技术调用PC的摄像设备定时对考生拍照,使用Web Socket技术实现图像数据实时传输,综合利用基于Ada Boost算法的人脸检测和基于Fisherfaces方法的人脸识别技术自动发现他人替考、多人作答、擅自离开、中途换人,及时保存证据并通知监考教师。通过仿真测试验证该系统的有效性,为在线考试的监控工作提供可行的辅助手段。Aiming at some conventional cheatings in online exam and the problems of current video monitoring method, presents the image based invigilation scheme. Designs and implements a real-time image based monitoring system for online exam. To realize taking photos regularly for examinees and images data real-time transmission, uses Web real-time technology such WebRTC and WebSocket. To auto discover the cheatings such as replaced by other, multiple answers, leave the seat, substitute examinee in midway, the AdaBoost based face detection and the Fisherfaces based face recognition are utilized comprehensively. Validates the feasibility of the system by simulation test.

关 键 词：考试监控 WebRTC WebSocket 人脸检测 人脸识别 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]