基于人工智能表情识别与差异分析的游戏化学习投入状态研究  

Research on Game-based Learning Engagement State Based on Artificial Intelligence Facial Expression Recognition and Differential Analysis

作　　者：张露 胡若楠 雷悦 尚俊杰[4] ZHANG Lu;HU Ruo-nan;LEI Yue;SHANG Jun-jie(School of Network Education,Beijing University of Posts and Telecommunications,Beijing,China 100876;School of Teacher Education,East China Normal University,Shanghai,China 200062;Institute of Education,University College London,London,United Kingdom,999020;Lab of Learning Sciences,Graduate School of Education,Peking University,Beijing,China 100871)

机构地区：[1]北京邮电大学网络教育学院,北京100876 [2]华东师范大学教师教育学院,上海200062 [3]伦敦大学学院教育学院,英国伦敦999020 [4]北京大学教育学院学习科学实验室,北京100871

出　　处：《现代教育技术》2023年第12期89-99,共11页Modern Educational Technology

基　　金：2023年度教育部人文社会科学研究规划青年基金项目“面向汉语二语语篇理解的在线虚拟情境学习体验的模型构建与评价研究”(项目编号:23YJCZH295)资助。

摘　　要：当前,游戏化学习凭借其增强学习动机的优势,受到教育领域的广泛关注,学习投入作为学生在游戏化学习中取得优异学业表现的关键因素,已成为当下研究者关注的重点方向,然而传统问卷调查方法在分析游戏化学习中的学习投入方面仍存在局限性。为了更加客观、准确、高效地探究游戏化学习中的学习投入状态,文章首先梳理了该领域的研究现状;然后,文章对学生的面部表情进行视频记录,并基于面部表情分析和K-means聚类方法,发现学生呈现了三种不同类型的情绪状态,即快乐学习型、愤怒焦虑型、认真谨慎型;最后,文章为探析不同情感投入类型下的差异性特征,对过程中的表情数据、游戏后台数据与学业成绩进行关联分析,发现快乐学习型和认真谨慎型的学生在分数概念性知识的后测中有较好的成绩表现。文章通过研究游戏化学习中的学生情绪状态类型及其差异特征,旨在为分析学习投入状态与认知表现的关系提供线索。At present,game-based learning has attracted wide attention in the educational field due to its advantages of enhancing learning motivation and improving academic performance.Learning engagement,as a key factor for students to achieve excellent academic performance in game-based learning,has become the focus of current researchers.However,traditional questionnaire survey methods still have limitations in analyzing learning engagement in game-based learning.In order to explore the learning engagement state in game-based learning more objectively,accurately and efficiently,this paper firstly reviewed the current research status in this field.Then,the paper made video recording of the students’facial expressions,and students’three distinct emotional states were identified,namely,joyful learning type,angry and anxious type,and diligent and cautious type based on facial expression analysis and K-means clustering.Finally,in order to explore the differential characteristics of these emotional engagement types,a correlation analysis was conducted between facial expression data and background game data in the process and academic performance.It was found that joyful learning type and diligent and cautious type of students performed better in the post-test of conceptual knowledge.By studying learners’emotional states of in game-based learning,this paper aimed to provide insights for analyzing the relationship between learning engagement and cognitive performance.

关 键 词：学习投入 游戏化学习 人工智能 表情识别 聚类分析 

分 类 号：G40-057[文化科学—教育学原理]