Application of time-frequency distributions to the independent component analysis of ECG signals

Abstract

This paper deals with the problem of independent component analysis of ECG signals. We propose to exploit the time frequency information present in the ECG signals. Existing approaches use only high and second order statistics to separate such kinds of signals. We investigate the application of time-frequency tools to perform the same task. This approach seems for us natural since the ECG signals show a special structure in the time-frequency domain of which we take advantage in our proposed analysis. This time-frequency analysis can contributes in the decision making of medical diagnoses. We propose to separate a combination of mother ECG signals and weak fetus ECG signals. Experimental results are reported and compared to results obtained with a higher order statistic method.