Real-time assessment of educational effectiveness using computer vision: A methodological framework demonstrated in a case study in Yerevan State University

This article presents a novel approach to measuring the effectiveness of learning processes in real time, based on computer vision emotion detection, thereby dispelling the need for traditional student feedback systems. Conventional evaluations conducted at the end of coursework often provide delayed and subjective feedback that does not accurately represent the learning process as it unfolds. Notably, our approach continuously captures and analyzes students' affective reactions during instruction and provides a real-time understanding of their level of engagement. Using the DeepFace library, we processed facial expression data obtained from 36 students of Yerevan State University (YSU) attending a video-based e-course on general psychology. Classification of expressions was primarily based on Ekman's (2014) theory, with the addition of the neutral emotional state to represent the model's special characteristics, which had been conceived as the absence of emotion (represented by a neutral facial expression). This model enables educators to identify concrete pedagogical examples linked with various affective responses, thereby ensuring a more subtle and empirically-based assessment of teaching effectiveness. Although the methodology is framed as a case study conducted in YSU, its usefulness extends well beyond this specific setting. By rendering certain technical modifications, the recommended approach can be economically applied in various learning environments, such as conventional face-to-face and non-computerized classrooms. The results provide a foundation for adaptive instructional design and promote the creation of learning environments that are not just emotionally supportive but also pedagogically sound. • The study presents a novel approach to measure real-time educational effectiveness. • Computer vision-based emotional recognition was employed. • A case study was conducted involving 36 students of the Yerevan State University. • The methodological approach is viable across various learning contexts. • This contributes to emotionally attentive and pedagogically sound learning systems.