基于面部表情识别对视频流中学生听课情绪进行分析  

Analyze the Emotion of Students in Video Stream Based on Facial Expression Recognition

作　　者：罗万艳 龚柯存 杨怡[1,2,3] 李昶 孙守康 唐东明 LUO Wanyan;GONG Kecun;YANG Yi;LI Chang;SUN Soukang;TANG Dongming(Southwest Minzu University,Chengdu 610041;School of Computer Science and Engineering,Southwest Minzu University,Chengdu 610041;General Office of the People's Government of Ningxia Hui Autonomous Region,Yinchuan 750000)

机构地区：[1]西南民族大学,成都610041 [2]西南民族大学计算机科学与工程学院,成都610041 [3]宁夏回族自治区人民政府办公厅,银川750000

出　　处：《现代计算机》2021年第18期117-121,共5页Modern Computer

基　　金：四川省科技计划资助(No.2019YFG0207)。

摘　　要：教学过程中教师通过观察学生面部表情可以了解学生的学习情绪从而做出相应的教学决策,但教师不能记录每个学生听课中表现出来的情绪变化,特别是在在线教育中存在情感缺失的问题,从而导致做出的决策稍显片面。本文提出将面部表情识别技术应用于学生学习情绪识别,首先收集学生在真实课堂中听课时的面部表情数据建立训练数据集,进而搭建相应的神经网络模型进行实验。最后在真实数据集上进行了多组和SVM模型的对比实验并应用于在线流媒体平台,实验结果表明本文提出的方法能获得较高的准确率。During the teaching process,teachers can observe the students’facial expressions to understand their learning emotions and make corre⁃sponding teaching decisions.However,teachers often cannot record the emotional changes shown by each student during the lecture,espe⁃cially in online education,there is a lack of emotions,which leads to slightly one-sided decisions.This paper proposes to apply facial expression recognition technology to students’learning emotion recognition.Firstly,facial expression data of students listening in real class is collected to establish a training data set,and then a corresponding neural network model is built for experiments.Finally,multiple sets of comparative experiments with SVM models are performed on a real data set and then applied to online streaming media platform.The experimental results show that the method proposed in this paper can obtain higher accuracy.

关 键 词：神经网络 学习情绪 表情识别 表情数据集 流媒体 

分 类 号：G632.4[文化科学—教育学] G434[自动化与计算机技术—计算机应用技术] TP391.41[自动化与计算机技术—计算机科学与技术]