Enhancing Online Teaching Effectiveness Through Computer Vision Analysis of Teacher Expressions and Gestures in Educational Videos

With the proliferation of online education, improving the interactivity and effectiveness of online teaching has become a pressing issue.Computer vision technology, with its powerful capabilities in video and image analysis, can be used to deeply analyze teachers' facial expressions and body movements in educational videos, thereby assessing their impact on teaching effectiveness.Although some studies have attempted to apply these techniques, most methods overlook the temporal and spatial features of facial expressions and movements, leading to insufficient recognition accuracy.This paper proposes two innovative methods: a facial expression recognition method for teachers based on facial action units and temporal attention, and a gesture recognition method based on spatiotemporal feature disentanglement.These methods can more accurately capture and analyze the dynamic expressions and movements of teachers, providing new technical support for online education, with the expectation of significantly improving online teaching effectiveness.