The Research On Electrocardiogram Quality Assessment And Heart Rate Variability Analysis

Heart rate variability(HRV)signal contains a generous amount of information on physiological and pathological information of cardiovascular system,which is widely used to quantitatively evaluate the autonomic nervous system function.As we know,Heart rate(HR)is the result of autonomic nervous system regulation,which can be seriously affected by cardiovascular diseases,pathological states and life stress.Extracting the HR signals and recognizing the abnormal HR information about diseases are important for early warning of cardiovascular diseases.However,in the electrocardiogram(ECG)collecting,ECG is inevitably contaminated with noise,causing extracting accuracy of HR signal to descend.Therefore,the quality assessment of ECG signal is an important part of HRV analysis.In this thesis,the quality assessment of ECG signal and HRV analysis would be further researched.The main contents of this thesis are as follows:1)For the quality assessment of Multi-lead ECG signal,a Multi-lead ECG fusion algorithm(NFDA)is proposed.In the algorithm,the idea of local weighted linear prediction algorithm(LWLPA)is used to convert the multi-lead ECG signals to a single physiological signal.For effectively reserving the quality characteristics of multi-lead ECG signals,fuzzy inference system(FIS)is reasonably designed to estimate the weighted coefficient in NFDA.Experimental results show that NFDA algorithm can obtain desirable results on synthetic ECG signals,noisy ECG signals and realistic ECG signals.In the thesis,the theories of phase space reconstruction and recurrence quantification analysis(RQA)methods are applied to extract the quality feature parameters of ECG signal.Three quality features,Recurrence Rate(RR),Determinism(DET)and Entropy(ENT),are chosen by analyzing the results of the experiments.2)For effective classification of ECG signal quality,a spectral clustering via local projection distance measure(LPDM)is proposed.In this thesis,the Local-Projection-Neighborhood(LPN)is defined,which is a region between a pair of data and other data in the LPN are projected onto the straight line among the data pairs.Utilizing the Euclidean distance between projective points,the local spatial structure of data can be well exploited to measure the similarity of objects.Then the affinity matrix can be obtained by using a new similarity measurement,which can squeeze or widen the projective distance with the different structure information of data.Experimental results show that the LPDM algorithm can obtain desirable results with high performance on synthetic datasets,real-world datasets and images.3)For detecting the R-peak in ECG,in this thesis,the improved ensemble local mean decomposition(IELMD)algorithm is proposed as a tool to adaptively decompose a signal into a collection of demodulating AM-FM signals.In ensemble local mean decomposition(ELMD)algorithm,the mode mixing problem may be caused by the lower resolution of signal and the useful information from AM-FM signals cannot be extracted.In this thesis,an improved ensemble local means decomposition based on cubic spline interpolation is proposed,which is a useful technique to interpolate between data points and improved the resolution of signal.The results show that the IELMD can effectively eliminate the mode mixing problem and extract useful information from ECG signal.Thus,IELMD algorithm is superior to the local mean decomposition(LMD)and ELMD.4)In this thesis,a new algorithm for detecting R-peaks in ECG based on IELMD is presented.In the algorithm,the R-peak detection algorithm process can be divided into three stages: signal preprocessing stage,feature extraction stage and decision making stage.The algorithm performance is evaluated by the MIT-BIH arrhythmia database and the results show that the proposed algorithm can effectively achieve the goal of detecting R-peaks of ECG signal accurately.5)For HRV analysis,the nonlinear dynamical properties of HRV signal cannot be adequately described by Poincare plot.In this thesis,three dimensions Poincare plot is presented in which the inter-relationship between consecutive points in Poincare plot is well employed and the descriptive capability of nonlinear system is further enhanced.The algorithm is applied to the RR interval series derived from healthy subjects,coronary heart disease subjects and congestive heart failure,respectively.Results show that the states of cardiovascular system of aforementioned three types of subjects can be fully described by three dimensions Poincare plot.