Human Body Pose Recognition System Based on Teaching Interaction

In view of the problems of the high time cost and low accuracy of manual supervision in traditional classroom teaching, this paper proposes a human body pose recognition system based on teaching interaction. The enhanced basic network (ResNext-101 + FPN) was used in Mask R-CNN to extract the features of the input images. Then based on the behavior analysis algorithm and face detection data, the behavior data of each student in the classroom were obtained. Moreover, the behavior data were applied to support multi-dimensional visualization. The experimental results show that the system can timely and effectively reflect the learning status of students, and help teachers accurately grasp the classroom learning state of students, so as to adjust teaching strategies in a targeted way and help improve the quality of teaching.