Home // ACHI 2020, The Thirteenth International Conference on Advances in Computer-Human Interactions // View article

Engagement Estimation for an E-Learning Environment Application

Authors:
Win Shwe Sin Khine
Shinobu Hasegawa
Kazunori Kotani

Keywords: E-learning; Engagement; Fine-tuning.

Abstract:
In this study, we conducted an estimation of engagement through virtual learning environment by using facial images. We aim to improve student learning rates and get a better understanding of them through facial expressions. Nowadays, computation power and memory capacity are available for analysis on large scale datasets. As a result, deep learning techniques can effectively extract useful features from the given dataset over traditional approaches. Unfortunately, deep learning-based methods require a massive amount of labeled data. Although there are many face datasets for face related problems, such as face detection and face recognition, it is still limited to facial expressions. To overcome this limitation, we use the advantages of the style transfer technique to obtain the basic features of the face and eliminate the features that are not useful for engagement estimation. In our experiment, we use the VGG-16 face model to extract the prominent basic features of the face and eliminates the non-related features by differing peak and neutral frames. We demonstrated the practical use of our method through the efficiency of detecting student engagement. The results show that our proposed method provides 50% accuracy in engagement estimation.

Pages: 1 to 6

Copyright: Copyright (c) IARIA, 2020

Publication date: March 22, 2020

Published in: conference

ISSN: 2308-4138

ISBN: 978-1-61208-761-0

Location: Valencia, Spain

Dates: from November 21, 2020 to November 25, 2020