Applied Research Of Human Identification Using ECG Signal Based On Deep Learning Model

Electrocardiogram can reflect the process of electrical activity of the heart,it has an important reference value in the study of heart basic function and pathology.With the development of biometric technology,many scholars at home and abroad have done more and more research on identification technology,especially in face recognition,eye film recognition and speech recognition,and it has reached the goal of commercialization.But in the current ECG identification research,there are still many problems to be solved.For example,the ECG signals,periods and intervals of patients with diseases are different from those of normal people.In order to solve the problems in the ECG signal,this dissertation proposes an identification method based on deep learning model.First of all,this dissertation used the PTB public database to preprocess the ECG signal.Because the original ECG signal contains a lot of noise,which affects the extraction of ECG signal features,in our work,the 4th order Butterworth band-pass filter was adopted to filter and eliminate the interference of noise in the ECG,and carried out the normalization processing at the same time.Then,three effective region extraction methods were proposed,which includes add windows for the ECG signal,R-point detection method based on deep learning and converse one-dimension ECG signal into two-dimension image.All of the above three methods are compared to select the best effective region extraction method.Finally,In this dissertation,the convolution neural network model in deep learning was used to study the identification,and the PTB ECG database was used to verify the performance of the method.The method in this dissertation was compared with the existing identification method based on ECG signal.In this dissertation,the convolution neural network model in deep learning is used to study the identification,and the PTB ECG database is used to verify the performance of the method.The method in this dissertation is compared with the existing identification method based on ECG signal.The experimental results show that the combination of ECG R-point detection method based on deep learning and ECG signal identification method based on deep learning has the best recognition results,which provides technical support for the development of ECG identification system with high accuracy,strong robustness and good anti-counterfeiting performance.