Using video analytics to give insights about events happening in classroom is a very important task in classroom surveillance systems. This paper proposes a new algorithmic frameworkto identify abrupt changes in a class room video and thenevaluate the attention level of students.The proposed algorithm is implemented with and without video key frame extraction approaches. The SSIM (Structural Similarity Index) approach for key frame extraction is used in this study. After extracting the key frames, the detection of face and upper body of the students to evaluate their attention level is performed on the key frames. The results comparing thealgorithms with and without SSIM reveals that the SSIM based algorithm gives better results. The algorithmic design of the proposed approach, the results obtained and sample cases are presented in this paper.
S. Ahmed, Krishnnan, N., Ganta, T., and Dr. Jeyakumar G., “A Video Analytics System for Class Room Surveillance Applications”, International Journal of Recent Technology and Engineering (IJRTE), vol. 7, no. 5S3, 2019.