The Research Of J Wave Signal Processingand Analysis Based On Multilayer NMF Algorithm

In electrocardiogram (ECG) signals， we usually call the J point which isbetween the electrocardiogram QRS complex and ST segment． J wave is like aspecial dome-shaped or spike type， when J point is away from thebaseline． This may lead to early repolarization syndrome， bundle branchblock， pericarditis， and even sudden death， J wave is getting more and moreimportant in the clinical diagnosis as a new index of the ECG of ventricularbipolar，but its signal often mixed in normal ST segment，using the traditionalelectrocardiograph， and diagnosed by experience cannot meet the practicalrequirements． This paper designs a virtual J-wave signal processing andanalysis system on the platform LabVIEW and Matlab．This paper adopted the virtual instrument programming languageLabVIEW to treat the mixing signals firstly， containing analyzing ECG signal，displaying wave form and heart rate， giving an alarm to abnormalphenomenon， storing wave form and playback． Filtering is by using digitalfilters and wavelet transform， which obtains satisfactory results． We can usethe time data of ECG data to control display time， ECG characteristic parameters and alarm system， also， it can store the data of wave form， andfinally generate a report that contains the main testing information， whichmakes the observe and diagnosis of ECG signal more direct and dynamic ratherthan restrict in easily stained electrocardiograph．Then， we extract the original J wave signal from the ST segment andanalyze the accuracy of extraction by multilayer NMF algorithm， showing thecharacteristics of hump shape J wave from the aspects of frequency domain，power spectrum， and spectral type， providing the basis for clinical diagnosisand increasing the reliability of the diagnosis of J wave． We estimate the powerspectrum of the J wave by different algorithms in AR model and find a mostaccurate power spectrum estimation-covariance method．Finally，we analysisJ wave in time domain and frequency domain using BMTK kit from theperspective of HRV and draw some parameters of J wave in the digitaldomain． It not only provides the basis for J wave clinical diagnosis but also thereliability of diagnosis．...