A new method for muscular visual fatigue detection using electrooculogram

Objective: Muscular visual fatigue (MVF)is increasingly common in clinic; However, there is no objective and effective means for the detection of muscular visual fatigue. This study focuses on a new method for muscular visual fatigue detection based on electrooculogram (EOG). Methods: We analyzed the mechanism that develops muscular visual fatigue and designed an experiment to induce muscular visual fatigue intentionally. And we recorded electrooculogram and critical fusion frequency (CFF) in the process. Then we got four electrooculogram physiological indicators and correlation between them and critical fusion frequency was analyzed. Finally, the indicators tendency, statistical difference and support vector machine (SVM) analysis were carried out. Results: The work shows that both wavelet packet barycenter frequency (WPBF) and average blink time (ABT) are significantly correlated with critical fusion frequency, tendency of both them has a good consistency, there is a significant difference for them both before and after muscular visual fatigue and that the trained support vector machine has a classification accuracy of 0.796 (SD 0.172) for states before and after muscular visual fatigue. Conclusion: Wavelet packet barycenter frequency and average blink time can be used for muscular visual fatigue detection, a certain degree of muscular visual fatigue occurred after induction and the trained support vector machine can achieve a good classification detection. We conclude that wavelet packet barycenter frequency and average blink time can be used for accurate muscular visual fatigue detection. Significance: This study is of great significance in muscular visual fatigue prevention and treatment.