Research On Denosing And Waveform Detection Of Electrocardiogram Based On Lifting Scheme

Cardiopathy, as a chronic disease, is the main threat to human life because the heart is the driving force of human blood circulation. As a noninvasive measurement, electrocardiogram, which can reveal the status of heart, is unreplaceable and valuable in both cardiopathy diagnosis and research.It is of significant value to research and develop home-use ECG monitoring and diagnosis system, which can provide signal collection, denosing, automatic diagnosis, and telemonitoring functions, fulfill people's need in pursuing higher standard living, shorten the distance between doctor and patient and meet the trend of modern ECG equipment development.In order to implement of this kind of equipment, the main topics of this paper focus on denosing algorithm, waveform detection method, and automatic dignosis software implementation.First of all, the ECG signal is too weak to get noised with respect to noises in data collection process. According to the denosing requirement of ECG signal and the drawbacks of the present denoising methods, a new denoisng algorithm based on lifting scheme is presented in this paper. The simulation results using MIT-BIH records show that the new method can perfectly remove high frequency noise, the complexity is low because of using lifting scheme, and processed signals have little distortion. Combining with wavelet decomposition and reconstruction method that used to remove the low frequency baseline drift, the processed ECG signal can fulfill the denosing requirement.Furthermore, ECG signal parameters can be obtained by applying wave detection algorithms. A modified QRS wave detection method is presented in this paper to localize the R-wave peak and calculate the width of the QRS wave. Lifting scheme is used to construct wavelet base which can improve the speed of computation. Some strategies are also presented to avoid false detection and omission and to raise detection accuracy. The detection methods of P wave and T wave are also introduced and simulated in this paper.Finally, based on the above algorithms, a windows application is developed using C#. The software is so powerful that provides signal denosing, waveform detection, automatic diagnosis, diagnosis report output, user history information management functions, etc. The key points of all these modules are introduced in this paper.Lifting scheme provides a new way to construction wavelet base in time domain, but has not widely used because theory system is not completion. In this paper, algorithms based on lifting scheme are used on ECG signal processing, which will extend the application and research domain of lifting scheme.