The use of computer vision techniques for the detection of the level of attention in remote classes has increased due to the Covid-19 pandemic, one of these techniques is head pose estimation, so the objective of this research is the use of this technique to detect the level of attention in remote classes. For this, a methodology has been used, this methodology consists of face detection, detection of facial landmarks, and head pose estimation. Technologies such as: Python, TensorFlow, OpenCV have also been used. A neural network has been trained for the detection of facial landmarks for best results. Finally, the results obtained show that head pose estimation can be used to detect the levels of attention in remote classes, this technique is affected by the number of participants and the image quality of the platform being analyzed. When evaluating the results through metrics such as the number of FPS, and the detection of the level of attention, satisfactory results were obtained, since the number of FPS has an average of 5.5 in the best scenario, and the level of attention corresponds with visual analysis of test videos.
CITATION STYLE
Pinzon-Gonzalez, J. G., & Barba-Guaman, L. (2022). Use of Head Position Estimation for Attention Level Detection in Remote Classrooms. In Lecture Notes in Networks and Systems (Vol. 358 LNNS, pp. 275–293). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-030-89906-6_20
Mendeley helps you to discover research relevant for your work.