A Video Analytics Based Solution for Detecting the Attention Level of the Students in Class Rooms

Classroom surveillance, using video cameras, affords enhanced understanding of student behavior. This paper proposes a new algorithmic framework to evaluate the attention level of students, from classroom videos. The live video of a class room, when a teacher is delivering the lecture, is the input to the proposed framework. This framework identifies the key frames from the video and then detects the attention level of a particular student. The paper perused the Structural Similarity Index Method (SSIM) to discern key frames in a video. Detection of drowsiness is then performed to deduce whether or not the student is sleepy. Scrutiny of facial expressions is carried out, to perceive the psychological state of the student in the classroom. Finally, detection of gaze is carried out to examine whether or not the student's attention is on the black board. The algorithmic design for the proposed approach, the results obtained and the sample test cases are presented in this paper.