Comparison of on-line wavelet analysis and reconstruction: with application to ECG

FEDERICO LESTUSSI; LEANDRO EZEQUIEL DI PERSIA; DIEGO HUMBERTO MILONE

Wuhan

Conferencia; 5th International Conference on Bioinformatics and Biomedical Engineering; 2011

IEEE

The wavelet transform is a consolidated tool forsignal analysis. However, it has limitations for processing onlineand long duration signals. In this paper we applied threealternatives for on-line processing based in the wavelet transformwhich are suitable for biomedical signals. The methods for onlinewavelet processing was compared with the off-line DWT asbaseline reference. The output for each methods was analyzedin detail by using the mean square error and the normalizedmaximum amplitude error. We concluded that the best performanceis obtain with Hamming windows and overlap-and-addfor reconstruction. As an application example, we show goodresults for the on-line processing of ECG signals.