Effects of Fatigue Detection With Adaptive Feedback on Sustained Alertness and Learning Outcomes in Video‐Based Learning

ABSTRACT Background Online video‐based learning often leads to fatigue, which detracts from engagement and learning outcomes. Previous studies have examined monitoring mental states like attention through electroencephalography (EEG) headsets, but limitations such as high costs, discomfort, and limited scalability persist. Objectives This study evaluates the effectiveness of facial recognition technology in detecting fatigue levels during video‐based learning. By using eyelid closure (PERCLOS) and mouth opening percentage (POM) indicators, it aims to provide adaptive feedback that supports engagement and reduces fatigue. Key research questions address the impact on learning outcomes, feedback accuracy, and technology acceptance across different learner groups. Methods Three groups were established in an experimental design: an experimental group receiving fatigue‐responsive feedback, a control group with random feedback, and a second control group with no feedback. Post‐experiment assessments measured learning outcomes, feedback accuracy, and technology acceptance. Results Findings reveal that adaptive, fatigue‐based feedback significantly enhances engagement and learning outcomes compared to random or no feedback. The experimental group maintained higher alertness in learning, reflected in both quantitative data and learner feedback. Conclusions Facial recognition technology offers a scalable and non‐intrusive solution to address fatigue in video‐based learning. Adaptive feedback based on real‐time fatigue detection improves learners' sustained focus, suggesting practical applications for future online education initiatives. Further research is recommended to optimise feedback mechanisms and explore long‐term impacts on learning efficacy.