Pulse Recognition of Cardiovascular Disease Patients Based on One-Dimensional Convolutional Neural Network

In the clinical diagnosis and treatment of traditional Chinese medicine (TCM), The classification of pulse is an important diagnostic method. The pulse signal of patients with cardiovascular disease is quite different from that of normal people. In pulse classification, traditional methods generally need to manually extract feature information of the time domain and frequency domain, and then perform pulse classification through machine learning methods such as KNN and SVM. However, the manually extracted features may not really represent the true features of the signal. Convolutional neural network (CNN) can automatically extract local features through different convolution kernels, and it can extract the original features of the signal very well. This paper designs a one-dimensional convolution (1D-CNN) residual neural network structure to identify the pulse signal. First, the original data is sliced and normalized, and then the data set is divided into training set and test set, Then take the processed pulse signal directly as input. This allows the computer to perform feature self-learning through the network structure, Finally, it is classified through the fully connected layer. The final classification accuracy rate can reach 97.14\(\%\), which is better than traditional machine learning classification methods such as SVM and KNN, It provides new ideas and methods for the classification and objectification of pulse signals.