Electrocardiogram (ECG) is a widely used non invasive method for diagnosis of cardiovascular diseases. While recording, baseline wander enters in ECG signals due to movement of electrodes and respiration. In order to extract clinical information with high accuracy, there is a need to pre-process ECG signal before diagnosis. In this paper, baseline wander removal algorithm based on Empirical Mode Decomposition (EMD) has been proposed. In order to remove baseline wander from ECG signals, proposed algorithm has been applied to ECG signals taken from MIT/BIH and CSE databases. Also, in order to examine the performance of proposed algorithm in real time situations, ECG signals recorded in laboratory have been considered as test signals. Power Spectral Density and signal power in frequency range 0-0.5Hz have been calculated for original signal and reconstructed signal.

baseline wander, ECG, EMD, IMF.