Analysis of Classroom Teaching Behaviors Based on Multimodal Data Model

Precise identification of classroom behavior can help teachers and students understand classrooms, and help promote the development of smart classrooms. This article designed a multi -mode data model supported by multimodal data support based on classroom teaching scenes, including classroom teaching layers, data collection layers, algorithm analysis layers, and application service layers. Analysis of classroom teaching behaviors, this article extracts the image characteristics in the video based on the deep learning algorithm of YOLO-V5, the voice recognition technology extracts voice characteristics. Multi -mode data model analysis has obtained good analysis results and recognition results. In order to verify the effectiveness of the selected model, the model performance was performed on the labeled classroom behavior data set. The test results show that the selected model shows good performance in the analysis and identification of classroom behavior in the education scene.