Sensor-based gesture recognition with convolutional neural networks

Abstract

Recently, sensor-based gesture recognition has shown its importance through its various practical uses in the fields of control systems, medical monitoring and judging abnormal behaviors. Deep learning has been widely implemented in sensor-based gesture recognition problems, and the results prove its effectiveness. In this article, we propose a recognition model based on convolutional neural networks (CNNs) and apply it to smartphone-based data. The average classification accuracy is better than that of traditional machine learning models, which is 69.24% on this dataset. Furthermore, we analyze the effect of different factors on the result of the CNN model.