SmartEmotionTutor: A Unified Framework for Emotion-Aware Adaptive Tutoring with Real-Time Face Recognition Attendance

Abstract

: Traditional Intelligent Tutoring Systems (ITSs) adapt to learner performance but often neglect affective and behavioral dimensions, limiting personalization. This paper introduces SmartEmotionTutor, a unified framework that integrates real-time emotion recognition, biometric attendance verification, and adaptive content delivery. The system incorporates a Negotiable Learner Model (NLM), enabling students to inspect and contest records of mastery, engagement, and attendance, thereby enhancing transparency and trust. A controlled study with 90 high school students compared SmartEmotionTutor with both traditional classroom instruction and a baseline ITS. Results indicate that SmartEmotionTutor achieved a 41% relative improvement in post-test scores, 97% attendance verification accuracy, and significantly higher engagement levels. These findings demonstrate the potential of combining cognitive, affective, and behavioral analytics with negotiable learner models to create trustworthy and scalable tutoring environments.