Artificial Intelligence Classroom Behaviour Monitoring AI Class Monitor

With a teacher-to-pupil ratio of 1:80 in most Nigerian schools, the quality of pupils’ attention and learning engagement continues to be on a nosedive trajectory with a negative impact on pupils’ motivation to learn, increasing drop-out rate, and the socio-economic risks of a poorly educated populace. According to a report by the United Nations Children's Fund, over 58.3% of pupils in Nigerian schools are not learning effectively (Vanguard News, 2018). Teachers are also becoming ineffective at inspiring pupils’ concentration via a classroom atmosphere that facilitates effective learning. The additional demands on teachers due to the higher pupil to teacher ratio also limits opportunities for teachers to offer personalised learning adapted to the children’s learning needs and providing the necessary psychological support.

This work explores how deep learning-powered cameras in a classroom can assist teachers to recognize and assess their pupils’ engagement levels and classroom behavioural patterns, with the goal of intervening when students need motivation and assistance. Specifically, the artificial intelligence algorithm assesses the students' facial expressions, classroom activities and their demonstrated levels of interest. We propose an integrated approach that includes neural network facial recognition and an Apriori-based mood extraction and human action detection to uniquely understand each pupil's learning needs. With a relatively low-cost camera that can quickly and precisely read pupils’ facial expression and class behaviour, many learning risk indicators such as emotional stress, personal frustrations, fatigue, academic disconnection and learning dysfunction can be automatically identified, allowing the teacher to intervene appropriately as needed.