ABSTRACT
Objective:To avoid the redundant computation based on the convolution operation in the traditional wavelet transform, and to remove the baseline wander noise existing in the course of collecting the ECG signal. Methods: Use the lifting wavelet transform with two wavelets, and constitute the ECG signal with the noise removed after decomposing, setting the subband coefficient including the noise to zero, and rebuilding. Results:Use MATLAB to remove the baseline wander noise in the ECG signal and bw provided by the MIT-BIH database, and the results show that the baseline wander was removed effectively. Conclusion:The baseline wander noise in the ECG signal can be removed accurately though the method mentioned above, the waveform information in the original ECG signal can be maintained effectively, and subsequently, that can provide help for detecting the characteristic parameters in the ECG signal.