Research on Biometric Identification Recognition Method of Radial Artery Combined with Convolutional Network

The radial artery of human body contains the physiological information of human body, which is nonstationary and non-periodic. In order to extract the characteristics of radial artery and improve the efficiency of human signs recognition. In this paper, a method based on CNN-EMD as radial artery feature vector was proposed to recognize human signs by support vector machine (SVM). Firstly, the original signal is decomposed by empirical mode to obtain the intrinsic mode function (IMF), and the information entropy of IMF is calculated. Secondly, convolutional neural network (CNN) is applied to the original signal to extract morphological features, and the result and information entropy constitute the feature vector. Finally, the support vector machine algorithm is used to classify the extracted features. A total of 219 radial artery samples from Creative Commons platform were used to identify normal and abnormal human features with an accuracy of 78.57%. The study shows that the proposed method improves by 8.57% compared with the commonly used method of EMD combined with SVM, which shows that the method can effectively identify human body features.