Construction of an Intelligent ClassroomTeaching Evaluation System IntegratingMultimodal Deep Learning

In the context of rapid development of educational informationization, traditional classroom teaching evaluation methods have become difficult to meet the needs of modern teaching quality improvement due to issues such as strong subjectivity, poor real-time performance, and single data dimensions.This paper constructs an intelligent classroom teaching evaluation system that integrates multimodal deep learning, based on models such as computer vision (ResNet+PoseEstimation), speech processing (CNN+LSTM), and natural language processing (BERT+Transformer).It comprehensively analyzes multimodal data such as students' facial expressions, speech emotions, and classroom speaking content, accurately quantifying student focus, classroom interaction index, and teaching quality.By implementing a hybrid fusion strategy that combines Early Fusion and Late Fusion, effective integration of different modal features is achieved.Experimental results show that this system can objectively, in real-time, and comprehensively feedback the classroom teaching process, providing data-driven personalized improvement suggestions for teachers.The research provides a feasible path for the construction of an intelligent classroom teaching evaluation system, with strong practical application value and promotion significance.