Research on Gait Detection Algorithm Based on Plantar Pressure

Abstract

In order to identify the relationship between dynamic plantar pressure information and hallux valgus, this paper proposes a CNN model to achieve a gait detection method based on plantar pressure. The plantar pressure data during walking convert into an image, and then the convolution neural network model in deep learning is used. Given enough input image data and expected classification results. By changing the parameters of CNN, the relationship between plantar pressure image and hallux valgus was established. Then final classification model is obtained. The result shows that the accuracy of the proposed CNN model for the diagnosis of hallux valgus achieve 90.56%, which can be used as a clinical assistant method for the diagnosis of hallux valgus. The experimental results verify the validity of the model.