Learner’s attention detection in connected smart classroom using internet of things and convolutional neural networks

Abstract

Detecting learner attention is an essential part of learning assessment. Consequently, it becomes an essential requirement for adaptive intelligent teaching systems, to identify specific needs and anticipate orientations. In this article, we propose a new model of a connected smart classroom, based on the internet of things, artificial intelligence and machine learning to detect in real time learners' attention and marking their presence during the execution of a teacher-assisted pedagogical activity, as well as to adapt the most suitable learning objects to these learners. The proposed model is based on head position, gaze direction, yawning and eye-state analysis as facial landmarks detected by cameras connected via the Bluetooth low energy network and transmitted to a developed convolutional neural network. In addition, a series of experiments have been conducted to evaluate the performance and efficiency of the model developed. The findings demonstrate that the model developed can be used to precisely capture the status of learners in the classroom in terms of attention and identification. In this way, these interesting findings can be used to adapt teaching activities to the individual needs of learners, and to identify areas where they have difficulties and needs extra help.