中国教育类核心期刊 CSSCI来源期刊 RCCSE中国权威学术期刊

杂志文章

您的位置:首页 >文章目录 >2020年第10期>

    多维度特征融合的教学行为智能分析模式构建

    Construction of Intelligent Analysis Model of Teaching Behavior Based on Multi-dimensional Feature Fusion

【作      者】:

赵 刚, 朱文娟, 胡碧灵, 夏 青, 刘 闪, 初 洁


【关 键 词 】:

教学行为; 多维度特征; 智能分析; 人工智能技术


【栏      目】:

理论探讨


【中文摘要】:

分析课堂教学行为是揭示课堂教学规律的一个重要途径。如何利用信息技术手段处理与分析累积的海量教学视频公开课中的教学行为,成为当前改革教学过程评价服务的热点之一。由于教学中教学行为具有教学性、有序性、关联性等固有教育特性,使得基于深度学习的视频分割与识别技术仍然无法有效理解教学场景和教学行为。因此,文章设计了包含“基于视听觉特征的教学行为分析编码系统、教学行为听觉特征识别、教学行为视觉特征识别、教学分析过程数据可视化呈现”四个核心要素的一种多维度特征融合的教学行为智能分析模式,提出了“视觉特征为主、听觉特征为主、融合特征为主”三条实践路径,以厘清教学行为智能分析要素的关系,并以视觉特征为主的实践路径初步分析了43节课堂教学视频,提取了时间维度上的教学行为视觉特征,以期为“一师一优课、一课一名师”“教师研修”等教学行为智能分析活动提供借鉴。


【英文摘要】:

Analyzing classroom teaching behavior is an important way to reveal the law of classroom teaching. How to use information technology to process and analyze the teaching behaviors in the massive teaching videos in open classes has become one of the hotspots in the evaluation service of the current teaching reform. However, due to the inherent educational characteristics of teaching behaviors in teaching, such as teaching, orderliness and relevance, the video segmentation and recognition technology based on deep learning still cannot effectively understand the teaching scene and teaching behavior. Therefore, this paper designs an intelligent analysis model of teaching behavior with multi-dimensional feature fusion, which includes four core elements of "teaching behavior analysis coding system based on visual and auditory features, teaching behavior auditory feature recognition, teaching behavior visual feature recognition, and teaching analysis process data visualization presentation". Furthermore, three practical approaches dominated by "visual feature, auditory feature and fusion feature" are proposed to clarify the relationship between intelligent analysis elements of teaching behavior. And 43 classroom teaching videos are preliminarily analyzed based on the practical approach dominated by visual features, and the visual features of teaching behaviors are extracted from the time dimension, so as to provide reference for intelligent analysis activities of teaching behaviors such as "one excellent course for every teacher, one excellent teacher for every course" and "teacher training".

网站首页 | 联系我们 | 意见建议| 杂志订阅| 管理部登录

版权所有 © 电化教育研究 CopyRight ©e-EDUCATION RESEARCH 2011-2018 电话:0931-7971823 地址:西北师范大学《电化教育研究》杂志社