基于Dlib模型的云自习智能检测和监管系统设计
Design of cloud self-studying intelligent detection and supervision system based on Dlib model
胡阳 1史培中 1蔡秋茹 1尹文怡1
作者信息
- 1. 江苏理工学院 计算机工程学院,江苏 常州 213001
- 折叠
摘要
云自习的发展对培养学生的自主学习能力、推动智慧教育的建设有重要意义.然而云自习的用户监管正面临一些困难,如目前已有的视频监控方法仍需由监督者使用完整过程录像观测用户的自习状态.此外,完整的视频监控数据上传云服务器后,有较大的隐私泄露风险.为解决以上问题,文章设计了一种基于Dlib模型的人脸关键点检测算法,通过视频流分析用户自习状态下脸部特征的变化,实现用户自习状态的智能识别检测.根据上述检测算法,文章实现了云自习智能检测和监管系统.该系统无需存储完整的视频数据进行回看,即可展现用户的自习状态变化和自习结果,充分保护用户的隐私.自习场景模拟下的实验数据表明,该系统的平均检测正确率达到 80%以上,可实时处理每秒20 帧以上的视频流,能够满足自习状态检测准确率和实时性的要求.
Abstract
The development of cloud self-study is of great significance for cultivating students'self-learning ability and promoting the construction of smart education.However,user supervision for cloud self-study is facing difficulties.The existing video surveillance methods still require the use of complete process video recordings to observe the user's self-study status.Moreover,there is a significant risk of privacy leakage when complete video surveillance data is uploaded to the cloud server.To address the aforementioned issues,this paper proposes a Dlib model based facial keypoint detection algorithm that analyzes the changes in human behavior characteristics during user self-study through video stream analysis,achieving intelligent recognition and detection of user self-study status.Based on the above detection algorithms,this article implements a cloud self-learning intelligent detection and supervision system.The system does not need to store complete video data for review,and can fully display the changes in the user's self-study status and self-study results,fully protecting the user's privacy.The experimental data under the simulation of self-study scenarios shows that the detection accuracy of the system reaches an average of over 80%,and it can process video streams of more than 20 frames per second in real-time,meeting the requirements of accuracy and real-time detection of self-study status.
关键词
云自习/行为检测/智能监管/图像处理/Dlib模型Key words
cloud self-studying/behavioral testing/intelligent supervision/image processing/Dlib model引用本文复制引用
基金项目
国家自然科学基金(61602216)
江苏省高校"青蓝工程"优秀青年骨干教师培养对象(KYQ22003)
省级大学生创新创业训练计划项目(重点项目)()
出版年
2024
NETL
NSTL
万方数据