WiFi human body posture estimation algorithm based on Performer-Unet

According to the Performer-Unet-based WiFi human body posture estimation algorithm, a human body activity video is collected, real posture labeling information containing human body skeleton point coordinates and CSI data are extracted and input into an artificial neural network for training and loss labeling, the artificial neural network is optimized by adopting a gradient descent method, and a model is obtained; the CSI data stream of the collected video is processed through the model, and the human body posture is accurately recognized. According to the method, the cross-modal technology is introduced into the human body posture recognition algorithm, the posture recognition algorithm based on WiFi is trained, the cost is low, the application range is wide, good privacy protection is achieved, the application range of posture estimation in multiple fields is greatly expanded, and the defects of traditional algorithm application are overcome. 本发明的基于Performer-Unet的WiFi人体姿态估计算法，采集人体活动视频，提取包含人体骨架点坐标的真实姿态标注信息和