A Machine Learning based Eye Tracking Framework to Detect Zoom Fatigue

被引：1

作者：

Patel, Anjuli ^{[1
]}

Stynes, Paul ^{[1
]}

Sahni, Anu ^{[1
]}

Mothersill, David ^{[2
]}

Pathak, Pramod ^{[3
]}

机构：

[1] Natl Coll Ireland, Sch Comp, Galway, Ireland

[2] Natl Coll Ireland, Galway, Ireland

[3] Technol Univ Dublin, Fac Digital & Data, Dublin, Ireland

来源：

CSEDU: PROCEEDINGS OF THE 14TH INTERNATIONAL CONFERENCE ON COMPUTER SUPPORTED EDUCATION - VOL 2 | 2022年

关键词：

Eye Tracker; Zoom Fatigue; Machine Learning; SVM; KNN; Ada-Boost; Logistic Regression; Decision Tree; MENTAL FATIGUE;

D O I：

10.5220/0011075800003182

中图分类号：

TP39 [计算机的应用];

学科分类号：

081203 ; 0835 ;

摘要：

Zoom Fatigue is a form of mental fatigue that occurs in online users with increased use of video conferencing. Mental fatigue can be detected using eye movements. However, detecting eye movements in online users is a challenge. This research proposes a Machine Learning based Eye Tracking Framework (MLETF) to detect zoom fatigue in online users by analysing the data collected by an eye tracker device and other influencing variables such as sleepiness and personality. An experiment was conducted with 31 online users wearing an eye tracker device while watching a lecture on Mobile Application Development. The online users were given an exam followed by a questionnaire. The first exam was based on the content of the video. The online users were then given a personality questionnaire. The results of the exam and the personality test were combined and used as an input to five machine learning algorithms namely, SVM, KNN, Decision Tree, Logistic Regression and Ada-Boost. Results of the five models are presented in this paper based on a confusion matrix. Results show promise for Ada-Boost for detecting Zoom fatigue in online users with an accuracy of 86%. This research demonstrates the feasibility of applying an eye-tracker device to identify zoom fatigue with online users of video conferencing.

引用

页码：187 / 195

页数：9

共 25 条

[1] Distinguishing mental attention states of humans via an EEG-based passive BCI using machine learning methods [J].