Deep Learning-Based Visual Attention and Learning Behavior Analysis in Interactive Educational Technology

This study proposes a deep learning model for visual capture and learning behavior analysis in interactive educational technology. The approach aims to enhance educational quality and support students’ well-rounded development by integrating artificial intelligence. The study combines high-precision camera and image recognition algorithm to capture students' expressions and actions in class in real time, which provides data support for analyzing learning behavior and emotional state. Deep learning models, especially Convolutional Neural Network (CNN) and Gated Recurrent Unit (GRU), extract image features and process time series data to identify individual differences such as learning habits, interest preferences and cognitive styles. In addition, the model also integrates attention mechanisms to enhance the attention to key behavioral characteristics. The research results show that the model has high accuracy in facial expression recognition and gesture capture and can effectively identify students' learning styles and interest preferences and generate personalized learning plans. Through the implementation of a personalized learning plan, students' average grades and classroom participation have been significantly improved, which concludes that the model has great potential to enhance the pertinence and effectiveness of teaching.