基于人脸表情特征的高校课堂教学质量在线评估模型  

An online evaluation model for classroom teaching quality in universities based on facial expression features

作　　者：张成叔[1,2] ZHANG Cheng-shu(School of Computer and Information Technology,Hefei University of Technology,Anhui Hefei 230600,China;School of Information Engineering,Anhui Vocational College of Finance and Trade,Anhui Hefei 230601,China)

机构地区：[1]合肥工业大学计算机与信息学院,安徽合肥230600 [2]安徽财贸职业学院信息工程学院,安徽合肥230601

出　　处：《齐齐哈尔大学学报（自然科学版）》2024年第3期11-15,共5页Journal of Qiqihar University(Natural Science Edition)

基　　金：2020年度安徽高校自然科学研究重点项目“基于深度视觉的室内移动机器人同步定位与建图研究”(KJ2020A1121);2021年安徽省高等学校自然科学研究重大项目“混合式教学环境下AI赋能自适应学习研究”(KJ2021ZD0175)。

摘　　要：针对高校课堂教学质量在线评估模型识别率较低,评估过程主观性较强的问题,提出基于人脸表情特征的高校课堂教学质量在线评估模型。提取高校课堂人脸表情特征,利用图像层、S1层、C1层、S2层和C2层进行特征匹配和评选,使用贝叶斯分类模型对特征图像的平滑参数进行优化,确定使用率先验概率,判断学生的状态,评估课堂质量。实验结果表明,提出评估模型的评估率优于传统评估模型,在10~30 min内,学生的听课率最高,因此可以将重点问题和难点问题在第10~30 min内讲解,提高教学质量。In response to the low recognition rate and strong subjectivity of the online evaluation model for classroom teaching quality in universities,a facial expression based online evaluation model for classroom teaching quality in universities is proposed.Extract facial expression features from university classroom classrooms,use image layers,S1 layers,C1 layers,S2 layers,and C2 layers for feature matching and selection.Use Bayesian classification models to optimize the smoothing parameters of feature images,determine a prior probability of usage,judge students'status,and evaluate classroom quality.The experimental results show that the evaluation rate of the proposed evaluation model is better than that of traditional evaluation models.Within 10 to 30 minutes,students have the highest listening rate.Therefore,key and difficult problems can be explained within 10 to 30 minutes to improve teaching quality.

关 键 词：人脸表情特征 贝叶斯分类 教学质量 质量在线评估 评估模型 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]