ABSTRACT
Wavelet transforms were used to remove noise and detect characteristic points for feline and human electrocardiograms. The electrocardiograms contained both 60 Hz line noise and wideband noise. Noise was removed from the signals using both Fourier and wavelet techniques. For the Fourier technique, the signals were notch filtered at 60 Hz, and then low-pass filtered. To remove noise using wavelet techniques, the signals were wavelet-transformed and then soft thresholding was performed on each of the resulting details. The two methods of removing noise were then compared to illustrate the advantages of the wavelet approach. Wavelet transforms were also used to detect a characteristic point on each beat. An algorithm was written to find this point, calculate the instantaneous heart rate, and then plot heart rate variability.
