A Study On The Algorithm For ECG Signal Processing Based On The Second-generation Wavelet Transform

Wavelet transform is a new tool for mathematic analysis arisen in the late of 1980s, overcoming the shortcomings of Fourier transform because of its good localization in both time and frequency domain. It has been widely applied in many fields. In 1996, Sweldens proposed lifting scheme, namely the second-generation wavelet transform, a new method used to construct wavelets and perform wavelet transform. This thesis focuses on the lifting scheme compares the traditional WT with it, discusses the lifting steps for biorthogonal spline wavelet then proposes the lifting scheme for bior1.3 and bior4.4.Based on these, further studies on the application of lifting scheme in the field of ECG signal processing were made.ECG signal contains abundant information about human heart, whether it can be detected well or not is important for clinical diagnosis. With development of the information technology, using computer to achieve automatic and accurate analysis for ECG has been a subject fervently pursued by both internal and external researchers. On the basis of predecessors' achievement, the following two aspects have been further studied:Firstly, the detection algorithm for QRS complexes: According to two-channel perfect reconstruction condition and WT's Multi-resolution Analysis (MRA) characteristics, a biorthogonal spline wavelet filter banks is designed and its lifting scheme is used to detect the R-peak, then the onset and offset of QRS complexes are confirmed farther. This method reduces the complexity, speeds up the programme's operation and achieves high detection rate of R-peaks. It is proved well by the American MIT database.Secondly, the compression of ECG signals: In this paper, the integer-integer WT based on db9/7 is used to process the ECG signals, the wavelet coefficients are afterwards encoded by SPIHT. The experiments on selected records from the MIT-BIH arrhythmia database indicate that the algorithm is easier and it can achieve better compression performance than the algorithm based on traditional WT.The method proposed in the paper can achieve the detection of QRS complexes and the compression for ECG signals through elementary experiments, which are valued experiences for the realization of practical application of the algorithm for ECG signal processing based on the second generation WT.