Seema
English
July 2025
This study explores how emotion recognition systems based on AI can work in educational classrooms to analyse the emotional state of students and their engagement levels. The AI implemented was able to collect real-time emotional reactions of students attending primary and middle schools based on facial expression analysis, eye-tracking, and the evaluation of voice tone of the students aged 8-14 years. The methodology was qualitative and experimental in nature and based on logs of AI-generated emotion and logs of teacher observation. The result showed that the most prevalent emotional state identified was Attentive and Happy which means that they were broadly engaging. Nevertheless, a large number of students demonstrated negative emotional states of Confused, Bored, and Frustrated especially during the second part of the classes. With a maximum of Attentive and minimum of Confused, they found that the match of AI-teacher emotion is going to be the most accurate, which indicates that both human and automation tools should be used because they enhance each other. It is concluded that the state of emotion recognition based on AI can be used as a valuable support tool of an educator in real-time, which can be used to make timely alterations in the pedagogy. Nevertheless, the moral aspect, error constraints, and situational meaning are still major elements of the prudent utilization of such technologies
199- 213
© Airo Journals 2021
