基于背景消减法的考场作弊行为检测方法

摘要

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.%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.