I. Introduction
With the rapid advancement of deep learning technology and the popularization of applications, its applications can be seen in computer vision, speech recognition, robot application, biomedicine, automotive electronics, meteorology, etc. Deep learning has advantages over traditional machine learning methods in classifying data. Features need to be extracted and classified when classifying data. These features are carefully analyzed. Using traditional machine learning methods to extract, a lot of useful information will be lost. Deep learning, by contrast, will capture. The feature selection process is automatically completed by the deep network, and more accurate features can be obtained through a higher amount of calculation to improve the recognition rate.