基于智能视觉的考试作弊行为识别仿真  

Recognition Simulation of Cheating on Exam Based on Intelligent Vision

作　　者：梅毅[1] 孙洪伟[1] 

机构地区：[1]南昌大学科学技术学院,江西南昌330029

出　　处：《计算机仿真》2014年第4期219-222,共4页Computer Simulation

摘　　要：研究考试作弊行为识别方法,由于在考试的过程中,通过考试现场的图像能够进行考试作弊行为识别。但是,考生可能由于非作弊原因导致身体局部发生非约束形变,从而对作弊行为识别带来干扰,从而导致误判情况的发生,最终降低了考试作弊行为识别的准确性。为提高准确率,提出了一种基于免疫多Agent算法的考试作弊行为识别方法。利用免疫多Agent方法,提取与考试作弊行为相关的特征,为考试作弊行为识别提供依据。根据上述考试作弊行为识别特征,按照作弊行为多衡量标准,完成考试作弊行为的识别。实验结果表明,利用本文算法进行考试作弊行为识别,能够提高识别的准确性,从而保证考试的公平性。The recognition method of cheating behavior was researched in this article to improve the accuracy of i dentification. In the process of examination, through the image recognition of examination site, the cheating behavior can be detected. However, the under constraint body deformation of non cheating candidates may cause interfer ence with the image identification, resulting in the occurrence of false positives, and ultimately reducing the accuracy of exam cheating recognition. To this end, we proposed an immune multi agent algorithm to make exam cheating recognition. Through this method, the cheating related features can be extracted and examined, which provides a ba sis for identifying exam cheating. According to the above identifying characteristics of exam cheating and several measurements of cheating behaviors, the exam cheating recognition was completed. The experimental results show that the use of this algorithm for exam cheating recognition, can improve the accuracy of recognition, so as to ensure the fairness of exams.

关 键 词：智能视觉 考试作弊 身体形变 考试公平性 

分 类 号：F127[经济管理—世界经济]