~随着近年来信息技术产业的飞速发展，家庭宽带网络和个人高速移动网络的快速普及，以及“终身学习”理念被人们广泛接受，在线教育应运而生并不断推广。2020年新型冠状病毒疫情爆发，响应疫情期间“停课不停学”的号召，在线教育瞬间成为全国各大高校和中小学以及各大教育机构的唯一选择。
人们对于在线学习已经并不陌生，甚至对于在线学习有了更高的要求，希望在线教师能够像传统课堂上的教师一样，关注到自己的学习状态，并能够根据自己的学习状态来调整授课进度和授课方式。现有的在线学习系统普遍侧重于知识点层面的交互，缺乏对情感的交互进行探索研究和应用，存在较为严重的“情感缺失”现象，教育者与学习者无法进行面对面的沟通，教育者无法捕捉到学习者当前学习情绪以及学习中的情感变化，进而无法得知学习者的学习状态及课堂体验，未能发挥情感因素的积极作用。
人工智能与在线教育相结合的智慧学习模式成为了教育信息化的新趋势，力求改善传统在线教育中的“情感缺失”现象。本文分析了在线直播教学系统存在的问题，提出了具有情感反馈功能的直播模型，并详述该模型的原理及流程。应用表情识别技术，在学习过程中监测和跟踪学习者表情，为教师提供直观的用户学习反馈数据，使教师能够得知课堂中学习者的参与度及学习体验，调整教学策略，利用学习者情绪情感的反馈和引导，营造一个智慧化的学习环境。
本文通过对深度学习理论的研究，着重分析了卷积神经网络的结构和特点，针对AlexNet、VGG和GoogLeNet等三个经典卷积神经网络模型进行了深入探讨，并提出了一种适用于在线教育情感识别的简单的卷积神经网络模型。利用本文提出的卷积神经网络模型实现了一个情感识别模块。该模块可以对当前学习者的面部表情进行识别，根据识别结果对该学习者当前的学习状态进行分类，并将分类结果进行记录和反馈，达到了教育理论和技术有效统一的目的，一定程度上缓解了“情感缺失”问题。
最后利用当前全国学校都在展开在线教学的情况下，将情感识别模块进行在线实验，在一对多的直播课堂中进行验证。本实验样本为22名2018级本科学生，在线实验结束后对22名学生进行问卷调查。通过每位学习者的在线学习状态时序图和调查问卷结果分析得出情感识别模块识别结果的准确率，实验结果表明本文设计的情感识别模块具有较高的识别结果准确率，具有一定的实用价值。
 

~With the rapid development of information technology in recent years, home broadband networks and personal high-speed mobile networks are rapidly popularize, the concept of "lifelong learning" have been widely accepted, and online education has emerged and become popular. In 2020, during the outbreak of the new coronavirus, in response to the call of “stop classes and non-stop schools” propose, online education instantly became the only choice for universities, primary schools, secondary schools, and educational institutions. 

People are not strangers to online learning, and even have higher requirements on it. We hope that online teachers can pay attention to learning status just like teachers do in traditional classrooms, and adjust the teaching progress and teaching methods according to learning status. Existing online learning systems generally focus on knowledge-level interactions, and lack of research and application of emotional interaction. Teachers and learners cannot communicate face-to-face, capture the learner's current learning emotions and the emotional changes, and then fail to know the learner's teaching effectiveness and classroom experience, and play the role of emotional factors.

The intelligent learning mode that combines artificial intelligence and online education has become a new trend of education informationization, and strives to improve the "emotional deficiency" phenomenon in traditional online education. This paper analyzes the problems of online live teaching system, proposes a live broadcast model with emotional feedback function, and details the principle and process of the model. Apply expression recognition technology detects and tracks learner expressions during the learning process, and provides teachers with intuitive user learning feedback data. It also enables teachers to gain student participation and learning experience in the classroom, adjusts teaching strategies, and uses student emotions and feedback Guide to create an intelligent learning environment

Through the study of deep learning theory, this paper focuses on the analysis of the structure and characteristics of convolutional neural networks, discusses three classic convolutional neural network models such as AlexNet, VGG, and GoogLeNet, and proposes a suitable strategy for online education, a simple convolutional neural network model for emotion recognition. Using the convolutional neural network model proposed in this paper, an emotion recognition module is implemented. This module can perform facial expression recognition on the face of the current learner, classify the learner's current learning status based on the recognition results, and record and feedback the classification results. It helps to achieve the purpose of effective unification of educational theory and technology, and a certain extent alleviates the problem of "emotional loss".

Finally, in the case of online teaching in schools across the country, the emotion recognition module is tested online and verified in a one-to-many live classroom. Samples of this experiment were 22 undergraduate students in the class of 2008. After the online experiment, 22 students were surveyed. The accuracy rate of the recognition result of the emotion recognition module is analyzed through the time sequence diagram of the online learning status from each learner and the results of the questionnaire. The experimental results show that the emotion recognition module designed in this paper has higher recognition results and certain practical value.

[1] 王红艳, 胡卫平. 中国在线学习研究现状与启示[J]. 中国远程教育, 2013(8):30-34,95.
[2] CNNIC. 中国互联网络发展状况统计报告[R]. 北京: 中国互联网络信息中心, 2019.
[3] 薛薇, 穆青, 聂阳阳. 青少年发展政策选编及评析(下)[M]. 北京: 北京理工大学出版社, 2012.
[4] 何克抗. e-Learning的本质——信息技术与学科课程的整合[J]. 电化教育研究, 2002, 1:3-6.
[5] 吴砥, 邢单霞, 蒋龙艳. 走中国特色教育信息化发展之路——《教育信息化 2.0行动计划》解读之三[J]. 电化教育研究, 2018, 39(06):32-34.
[6] Stylianos Asteriadis, Paraskevi Tzouveli, Kostas Karpouzis, et al. Estimation of behavioral user state based on eye gaze and head pose—application in an e-learning environment [J]. Multimedia Tools and Applications, 2009, 41:469-493.
[7] 李勇帆, 李里程. 情感计算在网络远程教育系统中的应用: 功能、研究现状及关键问题[J]. 现代远程教育系统, 2013(2):100-106.
[8] 徐振国. 智慧学习环境中学习画面的情感识别及其应用[D]. 济南:山东师范大学, 2019.
[9] 王济军, 马希荣, 何建芬. 现代远程教育中情感缺失的调查与对策研究[J]. 现代远程教育, 2007(4):29-31.
[10] 冯满堂, 马青玉, 王瑞杰. 基于人脸表情识别的智能网络教学系统研究[J]. 计算机技术与发展, 2011(6):194.
[11] 黄建, 李文书, 高玉娟. 人脸表情识别研究进展[J]. 计算机科学, 2016(43):123.
[12] 刘明远. 人脸表情识别算法研究与系统实现[D]. 呼和浩特: 内蒙古大学, 2016.
[13] Tsypes A, Burkhouse KL, Gibb BE. Classification of facial expressions of emotion and risk for suicidal ideation in children of depressed mothers: Evidence from cross-sectional and prospective analyses [J]. Journal of Affective Disorders, 2016, 197:147-150.
[14] 金辉, 高文. 基于特征流的面部表情运动分析及应用[J]. 软件学报, 2003,14(12):2098-2105.
[15] 马飞. 基于几何特征的表情识别研究[D]. 昆明: 云南师范大学, 2006.
[16] 杨国亮, 王志良, 王国江等. 基于非刚体运动光流算法的面部表情识别[J]. 计算机科学, 2007,3:213-229.
[17] 章品正, 徐琴珍, 王征. 基于神经网络树的面部表情自动分类方法[J]. 数据采集与处理, 2008,3:311-316.
[18] 刘姗姗, 王玲. 基于自动分割的局部Gabor小波人脸表情识别算法[J]. 计算机应用. 2009,11:3140-3043.
[19] 周慧. 人脸表情特征表达与识别算法研究[D]. 哈尔滨: 哈尔滨工业大学, 2010.
[20] 吴丹, 林学阎. 人脸表情视频数据库的设计与实现[J]. 计算机工程与应用, 2004,5:177~180.
[21] 张庆凯. 人脸表情视频数据库系统的实现及相关算法研究[D]. 沈阳: 东北大学, 2005
[22] Mahmoud Neji, Mohamed Ben Ammar. The integration of an emotional system in the intelligent system [J]. 2005 ACS/IEEE International Conference, 2005.
[23] RAY A, CHAKRABARTI A. Design and implementation of technology enabled affective learning using fusion of bio -physical and facial expression [J]. Educational technology & society，2016,19(4):112-125.
[24] BAHREINI K, NADOLSKI R, WESTERA W. Towards multimodal emotion recognition in e -learning environments [J]. Interactive learning environments, 2016,24(3):590-605.
[25] 孙波, 刘永娜, 陈玖冰等. 智慧学习环境中基于面部表情的情感分析[J]. 现代远程教育研究, 2015,2:96-103.
[26] 詹泽慧. 基于智能Agent 的远程学习者情感与认知识别模型——眼动追踪与表情识别技术支持下的耦合[J]. 现代远程教育研究, 2013,5:100-105.
[27] 马希荣, 王志良. 远程教育中和谐人机情感交互模型的研究[J]. 计算机科学，2005,32(9):182.
[28] 汪亭亭, 吴彦文, 艾学轶. 基于面部表情识别的学习疲劳识别和干预方法[J]. 计算机工程与设计, 2010,31(8):1764.
[29] 娄颜超. 智能化教学中的情感识别方法研究[D]. 武汉: 华中师范大学, 2011.
[30] 郑庆华, 刘均, 田峰等. 下一代e-Learning系统[J]. 中国计算机学会通讯, 2011,11:54-60.
[31] JAQUES P A, VICARI R M. A BDI approach to infer student's emotions in an intelligent learning environment [J]. Computers & education, 2007,49(2):360-384.
[32] CHEN C M, WANG H P. Using emotion recognition technology to assess the effects of different multimedia materials on learning emotion and performance [J]. Library & information science research, 2011,33(3):244-255.
[33] 乔向杰, 王志良, 王万森. 基于 OCC模型的 E-learning 系统情感建模[J]. 计算机科学, 2010,37(5):214-218.
[34] 马希荣, 刘琳, 桑婧. 基于情感计算的e—Learning系统建模[J]. 计算机科学, 2005,32(8).
[35] 徐晓青, 赵蔚, 刘红霞. 混合式学习环境下情绪分析应用与模型研究——基于元分析的视角[J]. 电化教育研究, 2018,39(8):70-77.
[36] 梁林梅, 李晓华. 让技术为学生提供更强大的参与经验——访哈佛大学学习技术专家克里斯·德迪博士[J]. 中国电化教育, 2010,(9):1-6.
[37] Jennifer A Fredricks, Phyllis C Blumenfeld, Alison H Paris. School engagement: Potential of the concept, state of the evidence [J]. American Educational Research Association, 2004,74(1):59-109.
[38] 刘斌, 张文兰, 江毓君. 在线课程学习体验：内涵、发展及影响因素[J]. 中国电化教育, 2016(10):90.
[39] 阴国恩, 李洪玉, 李幼穗. 非智力因素及其培养[M]. 杭州: 浙江人民出版社, 1998.
[40] 孔企平. 数学教学过程中的学生参与[M]. 上海: 华东师范大学出版社, 2003.
[41] G.西蒙斯. 网络时代的知识和学习——走向连通[M]. 上海:华东师范大学出版社, 2009.
[42] 文秋芳, 王立非. 中国英语学习策略实证研究20年[J]. 外国语言文学, 2004,(1):24-33.
[43] 李红梅, 刘宁, 李世改. 利用概念图减轻远程学习者认知负荷的策略[J]. 中国远程教育, 2007, 8:31-34.
[44] 刘志军, 冯永华. “顿覆论”下的慕课反思——兼论基于慕课的“课堂翻转”[J]. 课程.教材.教法, 2015(9):16.
[45] 赏维平. 师生的情感共鸣是课堂教学成功的要素[J]. 未来英才, 2017,(1):38．
[46] Ekman P, Friesen W V. Constants across cultures in the face and emotion [J]. Journal of Personality and Social Psychology, 1971,(2):124-129.
[47] A Mehrabian. Communication without words [J]. University of east London, 1968, 53-56.
[48] Ekman P, Friesen W V, Hager J. The Facial Action Coding System (FACS): A technique for the measurement of facial action [J]. Palo Alto, 1978.
[49] Carbonell J. Introduction: Paradigms for machine learning [J]. Artificial Intelligence, 1989,40(1):1-9.
[50] Dietterich T. Machine learning research: Four current directions (Final draft) [J]. AI Magazine, 1997,18(4):97-136.
[51] 刘建伟, 刘媛, 罗雄麟. 深度学习研究进展[J]. 计算机应用研究, 2014,31(7):1921-1930,1942.
[52] RUMELHART D E, HINTON G E, WILLIAMS R J. Learning representations by back-propagating errors [J]. Nature, 1986, 323(6088):533.
[53] HINTON G E, SALAKHUTDINOV R R. Reducing the dimensionality of data with neural networks [J]. Science, 2006,313(5786):504-507.
[54] LECUN Y, BOTTOU L, BENGIO Y, et al. Gradient-based learning applied to document recognition [J]. Proceedings of the IEEE, 1998, 86(11):2278-2324.
[55] Fukushima K. Neocognitron: A self-organizing neural network model for a mechanism of pattern recognition unaffected by shift in position [J]. Biological cybernetics, 1980,36(4):193-202.
[56] Hadji I, Wildes R P. What do we understand about convolutional networks [J]. arXiv, 2018,1803-1883.
[57] Akhtar S W, Rehman S, Akhtar M, et al. Improving the robustness of neural networks using K-support norm based adversarial training [J]. IEEE Access, 2016,4:9501-9511.
[58] Nair V, Hinton G E. Rectified linear units improve restricted boltzmann machines[C]. Proceedings of the 27th international conference on machine learning (ICML-10). 2010: 807-814.
[59] 苏文超. 人脸面部活动单元检测及微表情分析[D]. 北京: 北京邮电大学, 2019.
[60] 刘景福, 钟志贤. 网络教育的情感缺失现状及其对策[J]. 中国远程教育, 2001, 6: 15-17.
[61] 李施展, 朱家明. 基于卷积神经网络及几何特征对人脸的识别[J]. 上海工程技术大学学报, 2019,33(02):148-153.
[62] 汪雯琦, 高广阔. 基于PCA和SVM分类的跨年龄人脸识别[J]. 计算机时代, 2019,07:1-8.
[63] 魏刃佳. 基于情感识别的在线教学流程设计研究[D]. 西安: 陕西师范大学, 2014.
[64] 岳莹. 基于SVM的人脸检测与识别研究[D]. 石家庄: 河北科技大学, 2019.
[65] 杜成, 苏光大, 林行刚等. 改进的线段Hausdorff距离人脸识别方法[J]. 光电子·激光, 2005(01):89-93.
[66] 亢洁, 李佳伟, 杨思力. 基于域适应卷积神经网络的人脸表情识别[J]. 计算机工程, 2019,45(12):201-206.
[67] Krizhevsky A, Sutskever I, Hinton G E. Imagenet classification with deep convolutional neural networks[C]. Advances in neural information processing systems. 2012: 1097-1105.
[68] Szegedy C, Liu W, Jia Y, et al. Going deeper with convolutions[C] Proceedings of the IEEE conference on computer vision and pattern recognition. 2015: 1-9.
[69] Simonyan K, Zisserman A. Very deep convolutional networks for large-scale image recognition[J]. arXiv, 2014,1409-1556.
[70] Ashby F G, Isen A M. A neuropsychological theory of positive affect and its influence on cognition [J]. Psychological review, 1999,106(3): 529.
[71] 柴金焕, 马希荣. 基于情感计算的和谐人机教学模型的研究[J]. 微计算机信息, 2010(28): 219-221.
[72] 彭聃龄. 普通心理学[M]. 北京：北京师范大学出版社, 2010.
[73] 孙发勤, 邓雯心. 基于人工智能的课堂学习情感评价研究[J]. 中国教育信息化, 2019(23):58-62.
[74] 丁亦喆. 基于标签的个性化推荐方法研究[D]. 西安: 陕西师范大学, 2014.