A comparison of wavelet transforms through an HMM based ECG segmentation and classification system
BioMed'06 Proceedings of the 24th IASTED international conference on Biomedical engineering
Combining wavelet transform and hidden Markov models for ECG segmentation
EURASIP Journal on Applied Signal Processing
On the choice of filter bank parameters for wavelet-packet identification of dynamic systems
ICISP'10 Proceedings of the 4th international conference on Image and signal processing
| Hi-index | 0.00 |
Wavelet transform (WT) has attracted interest in applied mathematics for signal and image processing. In contrast to classical methods, such as the short-term Fourier analysis based on a finite windowed FFT, this new mathematical technique has been demonstrated to be fast and efficient in computation with localization and quick decay properties. Since 80's, WT has been proposed for signal processing in electrocardiogram (ECG) studies owing to its efficiency, large number of basis functions available, and high speed in data processing. It has been applied successfully in arrhythmia analysis, QRS complex detection, parameter extraction, data compression and smoothing. So far several papers have been published on applying WT in ECG signal processing. Indeed in this paper we address the problem of WT-based analysis of the ECG signal in order to improve the detection performance of various ECG profiles in actual noise conditions. The study is indeed concentrated on the comparison and selection of specific wavelets candidates able to enhance different sub-parts of the ECG beat signal (P, QRS and T waves). We also show how the different ECG sub-beat signals can be enhanced at different SNR levels and how it leads to propose a novel concept of composite wavelet transforms (CWT) using respective enhancement property of the two or three best candidates.