A Study On The Procession And Analysis Of ECG

Cardiovascular disease is one of the most important diseases in endangering human's life . and electrocardiogram(ECG) is a important diagnostic too! for assessing heart function. So the studies on ECG processing and ECG analysing are constantly followed with interest. With the progress of computer technologies , the methods of ECG processing and ECG analyzing have been improved .This paper approaches three questions about the study of ECG with the technology of computer.Firstly, in this paper a filtering method for ECG signal is introduced which decomposes original ECG signals in detail signals on different frequency bands by using Wavelet Transform(WT). According to ECG signal feature , some detail signals are processed by threshold filtering algorithm ,and then the signal is reconstructed by inverse WT . Therefore, several main interference in ECG signal can be removed. Experimental results demonstrate that this is an efficient filtering method.Secondly, a method of ECG's data compression based on wavelet neural network is introduced in the paper here. The Experimental results show that the wavelet neural network compressing template can greatly quicken the speed of network convergence. Comparing with BP network, the method of wavelet neural network has more advantages.Finally, a method of electrocardiogram classification based on wavelet neural network is reported in this paper. In this method , electrocardiogram data including those of P wave QRS complex and T wave were firstly used to train the wavelet neural network ,and then the wavelet neural network could correctly classify both the data used and not used in ECG training. Comparing the results of BP network , the wavelet neural network have better classification capability and generalization capability. Therefore , the wavelet neural network holds a bright future in ECG processing and ECG analyzing.