Infant physical fitness action standard identification and evaluation method based on convolutional neural network

The invention discloses an infant physical fitness action standard recognition and evaluation method based on a convolutional neural network, belongs to the technical field of artificial intelligence, computer vision and action recognition, and aims to solve the problems that existing infant physical fitness action evaluation depends on manual observation and recording, is low in efficiency and is easily influenced by subjective factors. Comprising the following steps: collecting video data of children during physical fitness training in real time, preprocessing the collected video data, extracting key point information of multiple persons in a video based on an OpenPose model, converting the key point information into a graph structure, processing the graph structure based on a Shift Graph convolutional network, capturing key nodes by using a TEA convolutional network, and outputting trajectory data. Constructing a 32-class body fitness action standard trajectory library, and comparing the output trajectory