Research on automatic identification of video classroom participation based on multimodal learning analysis
Data and Metadata · 12.05.2026 · 5 · 784
DOI: 10.56294/dm2026784 ↗Kurz gefasst
Fachartikel in der Zeitschrift Data and Metadata, online erschienen am 12.05.2026. Verfasst von Xuanzhen Geng und 1 weiteren Autor:innen. Die Arbeit nennt 18 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Introduction
Students' participation is one of the learning outcomes which measure the extent of their attention, interaction and cognitive engagement in the process of learning. The standard assessment methods which include manual observation and self-reporting need to be improved because they create bias and obtrusiveness while they become inefficient for handling large-scale video-based classroom settings.
Objective
The research aims to develop a robust-based multimodal learning framework for accurate real-time detection of student participation in video classrooms to solve existing shortcomings.
Method
The research used the Student Activity Video (SAV) dataset to create classroom video datasets, which include multi-view videos that show different student participation cues through their gaze direction and facial expressions and body movements. The videos underwent preprocessing through the application of noise reduction and normalization methods. The extraction of deep visual features proceeded through the implementation of convolutional neural networks (CNNs). A feature fusion technique integrated these features, followed by the development of the White Shark Optimized Enriched Capsule Neural Network (WSO-EnCapsNet) model for participation recognition.
Results
The proposed system outperformed existing methods, it achieved an accuracy rate of 97,87 percent and showed better performance in precision and recall and F1-score measurements. conclusions: The model functions as a strong and versatile tool which enables teachers to better their teaching methods while increasing student engagement through classroom participation assessment.
Zitieren
APA
Geng, X., & Xin, B. (2026). Research on automatic identification of video classroom participation based on multimodal learning analysis. Data and Metadata, 5, 784. https://doi.org/10.56294/dm2026784
Vancouver
Geng X, Xin B. Research on automatic identification of video classroom participation based on multimodal learning analysis. Data and Metadata. 2026;5:784. doi:10.56294/dm2026784
BibTeX
@article{Geng2026,
author = {Geng, Xuanzhen and Xin, Bangying},
title = {Research on automatic identification of video classroom participation based on multimodal learning analysis},
journal = {Data and Metadata},
year = {2026},
volume = {5},
pages = {784},
doi = {10.56294/dm2026784}
}
EUVIMED