The Research Of Arrhythmia Classification Based On ECG Signals

Because of the suddenness of cardiovascular diseases and the time-sensitive requirements for rescue work,it is important to prevent the occurrence of cardiovascular disease and pre-treatment,considering that patients often have corresponding arrhythmias before the onset of cardiovascular disease.Electrocardiogram(ECG)records changes in weak currents during heartbeat and is an important means of detecting abnormal types of heart rhythms.At present,the detection of abnormal heart rhythm is mainly based on the doctor's professional medical knowledge and work experience to identify and judge the ECG signal.However,the ECG signal data is more complicated and the data is large.The doctor needs to deal with a large number of ECG recognition every day,which is easy to cause errors.Judge.Therefore,the intelligent recognition processing of the ECG signals represented by the electrocardiogram has also become a research hotspot in recent years.In this paper,the denoising and classification algorithms of ECG signals are improved and optimized.Experiments are carried out on the well-known MIT-BIH arrhythmia database,and the results are compared with other methods.The main research contents and innovations of this paper are:(1)An ECG denoising method using a noise reduction algorithm in the EMD and wavelet domains is proposed,which can overcome the limitations of the existing methods.In the traditional EMD-based ECG denoising method,multiple eigenmode functions(IMFs)containing QRS complexes and noise are directly discarded.In order to preserve the QRS information in the presence of noise,the noise ECG signal is first enhanced by a window operation in the EMD domain.Then,the wavelet domain transform is performed on the ECG signal with relatively small noise.Finally,the adaptive threshold method is used to threshold the wavelet coefficients to reconstruct a clearer ECG signal.The simulation results show that compared with some prior art methods,the method can reduce the noise ofthe noise ECG signal more accurately and consistently.(2)A method based on deep neural network(DNN)is proposed to automatically classify abnormal ECG beats.DNN was developed using the Tensor Flow framework from Google's deep learning library,which consists of only seven hidden layers,with 5,10,30,50,30,10 and 5 neurons,respectively.The end result shows that our model is not only more efficient than the prior art in terms of accuracy,but also competitive in terms of sensitivity and specificity.