| Summary: | 碩士 === 國立高雄第一科技大學 === 電腦與通訊工程研究所 === 101 === Electrocardiogram (ECG) signal is a non-invasive method widely used for heart disease diagnosis. ECG can record the small potential change due to the cardiac contraction and the cardiac expansion. Doctors can track and treat the heart disease according to the ECG waveform of a patient.
1-Dimension discrete periodized wavelet transformation (1-D DPWT) has the word length growth effect and error propagation disadvantages, and 1-D DPWT can’t process signal in real time. K.C. Hung proposed a non-recursive discrete periodized wavelet transformation (NRDPWT) method. This NRDPWT can get all the wavelet coefficients of all levels in parallel. This NRDPWT improves 1-D DPWT disadvantages, and can process signal in real time. However, the 1-D NRDPWT generates artificial edge at boundary.
In this paper, a new method is proposed to reduce the artificial edge at the boundary. This new algorithm generates a new signal by flipping the input signal, and finds a new transformation matrix to overcome the artifact problem. This new method is called the new anti-artifact non-recursive discrete periodized wavelet transform (AA NRDPWT).
|
|---|
