Assessment of Visual Fatigue Caused by Eye-Controlled Interaction Based on Task Performance and Pupillary Response with GBDT-LR

Abstract

Highlights This study proposes a fusion of subjective and objective measures for visual fatigue induced in eye-controlled interaction by integrating eye movement features and task performance indicators. What are the main findings? A GBDT-LR model optimized via Bayesian optimization achieved an assessment accuracy of 89.79%. Its performance was comparable to that of the best-performing GBDT and substantially higher than those of LR, SVM, RF, and RF-SVM. Six non-redundant indicators spanning blink, fixation, pupil, and task performance dimensions were selected to characterize visual fatigue. What are the implications of the main findings? The lightweight pipeline provides a practical foundation for real-time fatigue monitoring and user experience optimization in eye-controlled systems. The method offers a non-invasive, easily deployable tool for assessing visual fatigue in eye-controlled interaction systems. The framework supports future adaptive interfaces that optimize user experience and promote sustainable use of eye-control technology. Abstract Assessing visual fatigue is crucial in eye-controlled interaction. Traditional methods are either overly subjective or rely on highly invasive, costly equipment and complex procedures that require expert supervision. This study proposes a machine-learning-based approach for visual fatigue assessment. Data collection employs non-intrusive, easily monitored eye-tracking to capture ocular eye movement data and task performance data, while subjective questionnaires label fatigue states. For feature selection, participant-level Wilcoxon signed-rank tests with Benjamini–Hochberg FDR correction were used to identify fatigue-related indicators, and a redundancy-removal step based on Spearman correlation yielded a final set of six non-redundant features. For the assessment method, we introduced a gradient boosting decision tree–logistic regression (GBDT-LR) model whose hyperparameters are optimized via Bayesian optimization. All models were evaluated under a unified 5-fold stratified cross-validation framework with within-fold standardization and nested hyperparameter tuning. Results indicate that this model can effectively predict the state of visual fatigue. Compared with the performance of five other models—gradient boosting decision tree (GBDT), logistic regression (LR), support vector machine (SVM), random forest (RF), and RF-SVM—the proposed GBDT-LR model achieved an assessment accuracy of 89.79%, demonstrating strong predictive performance. This study provides an effective method for predicting visual fatigue in eye-controlled interaction, laying a research foundation for optimizing the user experience of eye-controlled interaction and promoting the sustainable development of eye-control technology.