Student Emotion Recognition Using Computer Vision as an Assistive Technology for Education

Abstract

Research has shown that good educator or teacher empathy results in students with a greater understanding and acceptance, along with an environment that is conducive to learning. Educators that lack this attribute or are not cognizant of its value may potentially miss this opportunity, but the advent of affective computing methods has made automation of this task an interesting research avenue. This study explores the domain of education and provides an assistive technology that empowers the educator. It looks at integrating emotion recognition within a physical classroom setting to assist the educator with teaching. A model is proposed for achieving automatic student emotion recognition using computer vision methods to create an emotion report that is relevant to the educator. A prototype based on the model was successfully implemented that captures video, preprocesses it, isolates the relevant region of interest points in the scene containing students and classifies each captured student for each of the eight emotion classes, which is used to build a basic educator report. The preliminary results of the model show that deriving emotion from students in a physical classroom setting is feasible and can be achieved in near real-time while a class is being given. However, it is not without its limitations related to environment and equipment constraints, and further research needs to be done to determine how important emotion is in the learning process.