Enhancing Online Learning Monitoring with Novel Image Recognition Method Using Dlib for Eye Feature Detection

In the post-pandemic era, online learning has become a prevalent mode of education. To ensure effective monitoring of learners' engagement in online courses, this study proposes a novel method for real-time monitoring of learners' learning focus. It is divided into three main components: The first component is eye detection. We utilize Dlib technology and leverage pre-trained facial landmark detection to accurately identify the positions of the eyes. Through image processing and the application of appropriate thresholds (thresholds set above 190), this enables precise outlining of eye contours and recognition of eye movement characteristics, even during rapid eye movements or from varying angles. The second component is gaze estimation. By using adaptive thresholds for pupil segmentation and applying erosion and dilation to reduce image noise and enhance essential features, we can effectively determine pupil positions within images taken under varying photographic resolutions and lighting conditions. Angle calculation is subsequently employed to determine whether learners are concentrating on online learning courses. The third component involves an anti-cheating mechanism. By utilizing pupil-centered corneal reflex (PCCR) technology to address potential cheating behaviors of learners, an anti-cheating mechanism is proposed to detect and counter fraudulent tactics such as photos and makeup. Experimental results demonstrate that this study is capable of effectively monitoring learners' learning status and accurately determining whether learners exhibit appropriate learning attitudes and behaviors. This empowers a comprehensive assessment of learners' learning outcomes during the period of online learning.