The Analysis Of Pulseprint Timeliness Based On Convolutional Neural Network

With the development of economy and society,people put forward higher and higher requirements for information security and personal identification.Traditional identification methods such as keys,ID cards,passwords are increasingly difficult to meet people's requirements.Biometric technology that uses the uniqueness of human biometrics to identify personal identities has become the preferred method to ensure personal information security.Pulse is a palpable arterial pulse on the surface of the human body.Its formation is based on the contraction and relaxation of the heart and the elastic expansion of the arterial wall.Pulse wave contains abundant information of human physiology and pathology.Different people have different characteristics,and it is an important physiological signal of human body.Pulse,as a kind of biological characteristics,carries human identity information just like human fingerprints.With the progress of modem science and technology,the objectification of pulse diagnosis has been developed,and people use moderm methods to collect and study pulse signals.Convolutional neural network is an efficient recognition method developed in the field of machine learning in recent years.It has the advantages of strong robustness,easy training and optimization.So it is a powerful tool for classifying pulse characteristics.Based on the above background,the innovation and research contents of this paper are mainly divided into three parts:1.Optimize the original pulse analysis model based on convolution neural network,improve some shortcomings of the original model,and add the data preprocessing layer,preprocess the pulse data through SVD denoising,demodulation and de-mutation value.Experiments show that the improved model is superior to the original model.In the experiment of identifying people through pulse,the accuracy of the test set of the improved model reaches 92.09%,which is 8.28%higher than the original model.2.Pulse,as a kind of biological characteristics,also carries the identity information of human body.By analogizing fingerprints and voiceprint,this paper creatively puts forward the concept of pulse pattern,which enriches the existing biometric identification technology.On this basis,this paper analogizes the process of fingerprint recognition and voiceprint recognition,gives the process of pulse recognition,and points out the necessity of building a unified standard pulse database.3.With the optimized pulse analysis model as a tool,the timeliness of pulse is analyzed through experiments.The experiment proves that the pulse has time-effect,and the time limit of the pulse time-effect is more than or equal to 100 days.At the same time,this paper proposes that the multi-biometric recognition technology which combines different features and different recognition methods should be a direction of future development.