A Meta-Learning-Based Multi-Scene Student Posture Detection Method for Enhancing Learning Motivation and Engagement in Smart Classrooms

This study examines how smart classroom environments affect college students’ learning motivation and engagement, emphasizing the connection between the two. To support this, a meta-learning-based approach is proposed for recognizing student postures across different classroom settings. The methodology integrates a two-stage training process: an offline phase to build a posture detection metamodel using the MAML (ModelAgnostic Meta-Learning) framework, and an online phase for rapid adaptation to new environments using minimal labeled data. An adaptive optimizer is introduced to improve the model’s generalization and reduce overfitting in low-data scenarios. Experimental results demonstrate that the proposed method outperforms existing posture detection models in accuracy and adaptability across varied teaching environments. Students in smart classrooms also show significantly higher levels of engagement and independent learning compared to those in traditional classrooms. The study further finds a strong positive correlation between engagement and autonomous learning. These results highlight the method’s effectiveness in real-world classroom settings and its potential to enhance student-centered learning through intelligent posture analysis.