基于背景消减法的考场作弊行为检测方法  

Cheating behavior detection in examination room based on background subtraction

作　　者：蔺永政[1] 吴小羽 李恒建[1] 柳忠光[3] 

机构地区：[1]济南大学信息科学与工程学院,山东济南250022 [2]山东省计算中心,山东济南250101 [3]济南市公安局车辆管理所信息中心,山东济南250000

出　　处：《济南大学学报（自然科学版）》2015年第6期406-409,共4页Journal of University of Jinan(Science and Technology)

基　　金：国家自然科学基金(61303199);山东省教育科学规划(ZK1101322B021)

摘　　要：为了减少考场的作弊行为,保证考试的公平性,提出了一种将背景消减法应用于考场作弊行为检测的方法。通过分析交换物品的行为特点,提出了一种基于动态阈值的作弊行为检测算法:采用迭代阈值法确定动态阈值对差分图像进行分割,利用分割的结果指导背景的更新,得到良好稳定的背景模型,并基于背景减除算法完成了考场作弊行为的检测。实验结果表明,这种方法能够较好地检测出交换物品的作弊行为,并能克服光线等因素的影响,有效甄别考场的作弊行为。To reduce the number of cheating cases and assure the fairness in examinations,a kind of method for cheating behavior detection is proposed,in which the background subtraction technique is applied. By analyzing the behavioral characteristics of exchanging objects,a cheating detection method based on the dynamic threshold is presented. In the proposed scheme,differential image is segmented with the threshold determined by adopting iterative threshold methods. Then,the background updating is guided by segmenting results. Thus,an excellent background model is obtained. In addition,the detection of cheating in exams is fulfilled on the basis of background subtraction algorithm. Experiment results demonstrate that the proposed measure can find cheating behaviors of exchanging objects preferably and has the advantage of overcoming light influences. It can effectively identify examination cheating.

关 键 词：背景消减 作弊行为 动态阈值 行为检测 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]