Computer-Assisted Video Analysis Methods for Understanding Underrepresented Student Participation and Learning in Collaborative Learning Environments

试图了解课堂互动的研究通常依赖于课堂的视频记录，以便研究人员可以记录和分析教师和学生在学习环境中的行为。当研究是大规模的，这种分析是具有挑战性的，部分原因是它是耗时的审查和编码大量的视频。例如，在一个旨在提高拉丁裔/非裔学生计算思维和工程学习能力的课后项目中工作的学生之间数百小时的互动录像。这个项目正在探索使用计算机辅助方法进行视频分析，以支持研究人员手工编码。该项目正在调整用于医疗系统的计算机辅助诊断系统的程序。计算机辅助的过程可以生成摘要，然后研究人员可以使用这些摘要来识别关键事件，并描述课堂上的活动模式，例如学生之间的交谈或小组项目中的写作。创建摘要需要分析视频以进行面部识别、动作、颜色和物体识别。该项目将调查学生参与和教学的哪些部分可以使用计算机辅助视频分析进行分析。本项目由美国国家科学基金会EHR核心研究（ECR）项目、STEM+C项目和AISL项目支持。ECR项目强调在该领域产生基础知识的基础STEM教育研究。该项目由STEM+Computing项目资助，该项目旨在通过将计算思维和计算活动应用于STEM学科的教学和学习中，解决计算STEM领域的新挑战。作为加强非正式环境中学习的总体战略的一部分，推进非正式STEM学习（AISL）计划旨在推进非正式环境中STEM学习的设计和开发的新方法和基于证据的理解。这包括提供多种途径，扩大STEM学习经验的获取和参与，推进非正式环境中STEM学习的创新研究和评估，以及发展参与者对深度学习的理解。视频分析系统将为特定活动提供视频摘要，这将允许研究人员使用这些结果来量化学生的参与情况，并记录支持学生学习的教学实践。这将支持对大量视频数据的分析，而这些数据的分析通常很耗时。视频分析系统将识别场景中的物体，然后使用物体之间的距离测量和其他跟踪方法来编码不同的活动（例如，打字，交谈，学生和主持人之间的互动）。两组研究问题如下。(1)计算机辅助视频分析方法如何使数字视频的人工审查受益？视频摘要的哪些方面（例如，检测到的活动）可以帮助减少审查视频所需的时间？除了音频分析，未来在视频总结方面的哪些研究可以帮助减少回顾视频的时间？(2)我们如何使用计算机辅助视频分析方法量化学生的参与？计算机辅助视频分析方法可以准确测量学生参与的哪些方面？本研究使用的视频来自一个项目，该项目的重点是在课后环境中为拉丁裔/非拉丁裔学生提供工程和计算思维学习。数百小时的视频可供审查和分析，以设计和完善该系统。生成的编码还将有助于记录学习环境中的参与模式。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

Research that seeks to understand classroom interactions often relies on video recordings of classrooms so that researchers can document and analyze what teachers and students are doing in the learning environment. When studies are large scale, this analysis is challenging in part because it is time-consuming to review and code large quantities of video. For example, hundreds of hours of videotaped interaction between students working in an after-school program for advancing computational thinking and engineering learning for Latino/a students. This project is exploring the use of computer-assisted methods for video analysis to support manual coding by researchers. The project is adapting procedures used for computer-aided diagnosis systems for medical systems. The computer-assisted process creates summaries that can then be used by researchers to identify critical events and to describe patterns of activities in the classroom such as students talking to each other or writing during a small group project. Creating the summaries requires analyzing video for facial recognition, motion, color and object identification. The project will investigate what parts of student participation and teaching can be analyzed using computer-assisted video analysis. This project is supported by NSF's EHR Core Research (ECR) program, the STEM+C program and the AISL program. The ECR program emphasizes fundamental STEM education research that generates foundational knowledge in the field. The project is funded by the STEM+Computing program, which seeks to address emerging challenges in computational STEM areas through the applied integration of computational thinking and computing activities within disciplinary STEM teaching and learning in early childhood education through high school (preK-12). As part of its overall strategy to enhance learning in informal environments, the Advancing Informal STEM Learning (AISL) program seeks to advance new approaches to, and evidence-based understanding of, the design and development of STEM learning in informal environments. This includes providing multiple pathways for broadening access to and engagement in STEM learning experiences, advancing innovative research on and assessment of STEM learning in informal environments, and developing understandings of deeper learning by participants.The video analysis systems will provide video summarizations for specific activities which will allow researchers to use these results to quantify student participation and document teaching practices that support student learning. This will support the analysis of large volumes of video data that are often time-consuming to analyze. The video analysis system will identify objects in the scene and then use measures of distances between objects and other tracking methods to code different activities (e.g., typing, talking, interaction between the student and a facilitator). The two groups of research questions are as follows. (1) How can human review of digital videos benefit from computer-assisted video analysis methods? Which aspects of video summarization (e.g., detected activities) can help reduce the time it takes to review the videos? Beyond audio analytics, what types of future research in video summarization can help reduce the time that it takes to review videos? (2) How can we quantify student participation using computer-assisted video analysis methods? What aspects of student participation can be accurately measures by computer-assisted video analysis methods? The video to be used for this study is drawn from a project focused on engineering and computational thinking learning for Latino/a students in an after-school setting. Hundreds of hours of video are available to be reviewed and analyzed to design and refine the system. The resulting coding will also help document patterns of engagement in the learning environment.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.