| 基于卷积神经网络的考场不当行为识别 |
| |
| 引用本文： | 窦刚,刘荣华,范诚.基于卷积神经网络的考场不当行为识别[J].中国考试,2021(2). |
| |
| 作者姓名： | 窦刚 刘荣华 范诚 |
| |
| 作者单位： | 云南大学;昆明理工大学 |
| |
| 摘 要： | 将人工智能技术用于监控和识别考场中考生的不当行为,可以减轻监考人员的压力,提高考试的有效性、公平性和严肃性。本研究提出考场不当行为自动识别方案设想,以YOLOv3算法为核心,使用模拟考试场景视频数据开展自动识别实验,对考场不当行为自动识别的可行性和可靠性进行了检验。结果表明,考场不当行为识别的准确率高、速度快,方案可行且可靠,识别效果达到应用要求,对推动我国考试管理的智能化发展具有重要意义。
|
| 关 键 词： | 考试管理 考场不当行为 人工智能 标准化考场 |
Automatic Recognition of Misconduct in the Examination Based on Deep Convolutional Neural Network |
| |
| DOU Gang,LIU Ronghua,FAN Cheng.Automatic Recognition of Misconduct in the Examination Based on Deep Convolutional Neural Network[J].China Examinations,2021(2). |
| |
| Authors: | DOU Gang LIU Ronghua FAN Cheng |
| |
| Institution: | (Yunnan University,Kunming 650091,China;Kunming University of Science and Technology,Kunming 650504,China) |
| |
| Abstract: | It is helpful to improve the effectiveness justice and authority of the examination by applying artificial intelligence to monitor and recognize the examinees’misconduct,apart from reducing the pressure of invigilators.This paper proposes an automatic recognition scheme for misconduct in the examination room.With the YOLOv3 algorithm as the core,the experiment of automatic recognition is carried out on a simulated examination scenes to test the feasibility and reliability of the technology of automatic recognition in the examination room.The results show that the experiment boasts high recognition accuracy,good time effectiveness,viability and reliability,and meets the requirements for application,so it is significant to promote the intelligent development of examination management in China. |
| |
| Keywords: | examination management misconduct in examination room artificial intelligence standardized examination room |
| 本文献已被 维普 等数据库收录！ |
|
