Deep-learning-based image recognition for interactive digital classrooms

Abstract

With the rapid development of information technology, digital classroom has become an important part of education modernization. However, traditional human-computer interaction struggles to effectively capture student status and classroom dynamics, which affects teaching quality and personalized teaching promotion. Image recognition technology, especially deep learning-based intelligent visual algorithms, provides a new solution to improve classroom interaction experience. Based on CNN, YOLOv5, OpenPose and EfficientNet deep learning models and frameworks, this paper explores the application mechanism and implementation path of image recognition in digital classroom interaction. It systematically analyzes three core visual recognition tasks: teacher behavior recognition, student facial expression detection and student concentration detection, as well as intelligent processing of teaching resources. A complete digital classroom interaction system architecture integrating image recognition is proposed, and its effectiveness in improving student participation and teaching feedback efficiency is verified through empirical cases. The results show that image recognition technology integrating deep learning can significantly improve the intelligence of classroom interaction, with good scalability and practical application value, and plays a positive role in promoting the development of future smart education.