Classifying Student Attention Using Machine Learning To Develop Personalized Learning Systems

Many research has been undertaken to evaluate the attention span of students in educational settings, with the majority of them relying on qualitative approaches rather than a quantitative method to detect and assess attention . Quantitative approaches for evaluating pupils' attention have also been examined by several academics. Biometric wristbands were investigated as a measure of pupils' attentiveness . Students' attention to computer networking sessions was also inferred via facial expressions . Students' attention span was also determined using eye and head posture tracking . Teachers may be able to determine which education technique a specific student answers best to by categorizing students as attentive or indifferent.

Customized e-learning systems have been created using genetic algorithms (GA) and case-based inference (CBR) . Adaptive user interfaces have also been advanced as a result of personalized learning . These methods are intended for use in online learning settings. The approach described in this paper may be utilized in both online and traditional classrooms.

The purpose of this study is to describe a system that employs an RGB-D camera to track, calculate, and sign-up student motions, attitudes, facial expressions, and voice expressions in order to yield data that may be used to quantify student attention. Following that, the data is gathered, classified, and analyzed using machine learning algorithms in order to assess whether or not the pupils are paying attention. This system is required for the construction of the recommended personalized learning system.