A Deep Learning Approach for Automatic Detection of Learner Engagement in Educational Context

Asma Ayari, Asma Ayari, Mariem Chaabouni, Henda Ben Ghezala

2025

Abstract

The rise of online learning modalities, including fully online, hybrid, hy-flex, blended, synchronous, and asynchronous formats, has transformed educational landscapes. However, assessing learners' engagement in these environments, where direct teacher-student interaction is limited, poses a significant challenge for educators. In this context, Artificial Intelligence (AI) emerges as a transformative force within education, leveraging advanced algorithms and data analysis to personalize learning experiences and enhance teaching methodologies. The detection of engagement in educational settings is critical for evaluating the effectiveness of instructional strategies and fostering student participation. This study presents an approach to assess the learner’s engagement through detecting the facial emotions and classifying the level of engagement. This approach proposes a deep learning model specifically designed for automatic engagement detection in educational environments, employing a Convolutional Neural Network (CNN) approach. We introduce an optimized CNN model tailored for recognizing learner engagement through facial expressions in distance learning contexts. By integrating the foundational elements of traditional CNN architectures with the widely acclaimed VGG-16 model, our approach harnesses their strengths to achieve exceptional performance. Rigorous training, testing, and evaluation on an augmented dataset demonstrate the efficacy of our model, which significantly surpasses existing methodologies in engagement recognition tasks. Notably, our approach achieves an accuracy of 94.10% with a loss rate of 10.39%, underscoring its potential to enhance the assessment of learner engagement in online education.

Paper Citation

in Harvard Style

Ayari A., Chaabouni M. and Ben Ghezala H. (2025). A Deep Learning Approach for Automatic Detection of Learner Engagement in Educational Context. In Proceedings of the 17th International Conference on Computer Supported Education - Volume 1: CSEDU; ISBN 978-989-758-746-7, SciTePress, pages 372-379. DOI: 10.5220/0013283200003932

in Bibtex Style

@conference{csedu25,
author={Asma Ayari and Mariem Chaabouni and Henda Ben Ghezala},
title={A Deep Learning Approach for Automatic Detection of Learner Engagement in Educational Context},
booktitle={Proceedings of the 17th International Conference on Computer Supported Education - Volume 1: CSEDU},
year={2025},
pages={372-379},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013283200003932},
isbn={978-989-758-746-7},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 17th International Conference on Computer Supported Education - Volume 1: CSEDU
TI - A Deep Learning Approach for Automatic Detection of Learner Engagement in Educational Context
SN - 978-989-758-746-7
AU - Ayari A.
AU - Chaabouni M.
AU - Ben Ghezala H.
PY - 2025
SP - 372
EP - 379
DO - 10.5220/0013283200003932
PB - SciTePress