UNIVERSAL AND ADAPTIVE THRESHOLDING METHODS FOR MULTILEVEL DENOISING OF ECG SIGNALS USING ORTHOGONAL WAVELETS

In signal processing, discrete wavelets transform (DWT) is used to compress and denoise n-Dimensional signals. DWT can also be used to remove noises such as instrumentation noises, power line interference, noise from muscle artefacts etc., from the captured biological signals like electro-cardiogram (ECG). Performing wavelet transforms on the ECG signals is like passing the signal through bandpass filters. In DWT, several wavelets have been generated for this purpose. In this work, an arrhythmic ECG signal noised with a sinewave was denoised using the orthogonal wavelets like Haar, Daubechies, Coiflets and Symlets, with adaptive and universal thresholding algorithms. The diagnostic validity of the mathematical transformation of the denoised signal is assessed computationally as well as clinically.