Web Application to Track Student Attentiveness during Online Class using CNN and Eye Aspect Ratio

ABSTRACT

During this COVID-19 pandemic online class platforms are the only solution to transfer the knowledge in the field of education. Even though the physical classes are being practiced slowly in some countries, still academicians are in the need of online classes. In addition to content delivery, teachers are in the need to concern about throughout the class time whether the students are listening and be active in online classes. Due to more bandwidth consumption of the audio and video streaming, students can't be compelled to unmute the audio and video when the teacher delivers the content. So, there is no option for the teachers to observe the student's activity. With the advancement of technology and enhanced image analysis capacity of deep learning techniques, a system is proposed to compute the student's activity and can report it to the teachers during the class time itself. Drowsiness detection is tested using CNN based segmentation on our own set of 5000 images collected from 1000 students. The observed result shows 90% accuracy in predicting the drowsiness of the student by observing the face pattern of the student without streaming the video to the teacher's device. © 2022 IEEE.