Research On Quality Assessment Method Of Wrisband ECG Signal

In the 21st century,people's quality of life is getting better and better.However,different kinds of negative effects have come along with their high quality of life.The pressure of work and study,bad eating habits,and bad lifestyle have all seriously affected people's health.Cardiovascular disease is the most common chronic disease in recent days,which has the characteristics of high rate of sudden onset,high morbidity,high disability rate,and high mortality.Observation of ECG is the only way to analyze and diagnose heart disease at present.Therefore,in order to facilitate the acquisition of ECG signals,methods of collecting ECG signals through portable tools such as Mi Band and Apple Iwatch have emerged.However,ECG signals collected through wristbands and watches will be affected by factors like breathing,doing sports,sweating and changing environment,which will cause unnecessary noise interference to the ECG signals?motion artifacts,EMG interference,baseline drift?.These noises will not only result in equipment misdiagnosis and doctor misdiagnosis,but also increase the false alarm rate of cardiac monitors.In order to reduce the impact of these noises,it is important to know how to evaluate the quality of ECG signals and screen out the ECG signals with good quality.This paper will mainly study the quality assessment method of ECG signals.The content of this paper is as followed:1.According to the experiment,The frame division standard for the quality evaluation of wristband ECG signals is proposed,and the signal framing is done according to it.The research background of domestic and international ECG signal quality assessment is summarized,and the advantages and disadvantages of existing research methods are analyzed.2.on the basis of the existing ECG signal quality ecaluation method based on entropy measure^[14],an improvement is made,a method of ECG signal quality assessment based on wavelet packet decomposition-approximate entropy is proposed.Wavelet packet analysis is the most common time-frequency analysis method nowadays.According to the main energy frequency range of ECG signal and the energy frequency range of QRS wave group,the ECG signal is decomposed into three layers by wavelet packet decomposition.Eight nodes of the third layer are selected and reconstructed.The reconstructed approximate entropy is calculated,and the results are taken as feature vectors.Finally,the feature vectors are combined with SVM to realize the ECG signal quality assessment.3.Because the process of the wavelet packet decomposition of ECG signals depends on the choice of wavelet basis function,and illegal value will appear when calculating approximate entropy,a method of ECG signal quality assessment based on Empirical Mode Decomposition-sample entropy is proposed to solve this problem.Empirical Mode Decomposition?EMD?is an adaptive signal decomposition method,which does not depend on the choice of basis function and decomposition scale.Through Empirical Mode Decomposition of ECG signal,the IMF component of the signal is obtained,and the sample entropy of the IMF component of the signal is calculated,and the result is taken as feature vectors.Finally,the feature vectors are combined with SVM to realize the ECG signal quality assessment.4.Both methods mentioned above have some shortcomings in the characteristics of response signals.To improve these two methods,ECG signal quality assessment method based on feature fusion is proposed.Firstly,ECG signals are decomposed into three layers by wavelet packet decomposition.The wavelet packet coefficients of eight nodes in the third layer are calculated and the multi-scale entropy values of the eight wavelet packet coefficients are obtained as the first set of eigenvectors.Then,the local mean decomposition?LMD?of ECG signals is performed to calculate the multi-scale entropy values of decomposed PF components and the results are taken as the second set of eigenvectors.The two sets of feature vectors are fused to get the final feature vectors.Finally,the feature vectors are combined with SVM to realize the ECG signal quality assessment.