Classification and Recognition of Aerobics Movements Based on Convolutional Neural Networks

As a popular fitness activity, calisthenics has been paid more and more attention in modern society. This study discusses the classification and recognition of aerobics movements, analyzes the complexity of movements, lighting and occlusion problems, and difficulties in data set construction in this field, and constructs a movement recognition model based on convolutional neural networks by selecting appropriate network architecture, setting hyperparameters, formulating training strategies, and applying loss functions. The experimental results show that the proposed model is superior to the traditional model in the accuracy, speed and stability of action classification.