A novel approach for denoising the physiological signals using spectral methods

The electrocardiogram (ECG) is frequently used signal to diagnose heart problems. The interpretation and detection of physiological and pathological occurrences depends on high-quality ECGs. The recordings of ECGs, usually contaminated by high-frequency noise-components that are brought on by perturbations of the human body either internally or externally. These components severely effect the vital information of signals, therefore must be eliminated to recover the actual ECGs. In this study, a new method threshold filtered spectral method for ECG denoising is presented. This technique significantly reduces signal distortion while effectively removing high-frequency noise. It is based on thresholding and filtering the expansion coefficients. The validation of the proposed method is evaluated through the simulation results on MIT-BIH database. The experimental analysis demonstrates that the method yields satisfactory results.