Abstract
Digital smart classrooms are becoming increasingly popular. Many digital IoT devices have been integrated in smart classrooms in recent years to monitor and analyze students' behaviour and engagement levels, which greatly affects the environment of the smart classrooms. Student engagement is considered important for effective classroom learning but watching students' behaviour and keeping them engaged in a classroom is a difficult task. This paper provides a convolutional neural network architecture (CNN) for the analysis of students' behavioural engagement in real time in a smart classroom that exclusively employs only facial images from the input data. The proposed model uses a pre-trained VGG16 network because of the small amount of data available in the dataset. The pre-trained VGG16 network is modified and fed with our dataset for training and testing purposes. In particular, Internet of Things (IoT) technology is used in the smart classroom for collecting data related to students' behaviour. Real time analysis is done on the collected data using the fog-cloud computing platform to assess students' engagement in classroom. The test accuracy of 0.93 was achieved using the proposed model.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call