A New Physical Posture Recognition Method Based on Feature Complement-oriented Convolutional Neural Network

Accurate physical posture recognition is very necessary in high-level training and critical decisions of major events. The existing posture recognition algorithms cannot reflect the dynamic characteristics of athletes' postures. Therefore, this paper proposes a new physical posture recognition method based on feature complement-oriented convolutional neural network (CNN). This method extracts the global temporal and spatial features of physical posture to complete the feature complement. And a double-channel CNN posture recognition model is established. Deep learning is executed for physical image and energy physical history image by spatial channel and global time domain channel respectively. Then the features obtained by the two channels are fused to recognize the physical posture. Finally, experiments show that the proposed method has higher recognition accuracy than traditional methods.