Baseline Correction in Electrocardiographic Signals Using Dynamic Temporal Sequence Alignment

Abstract

this paper proposes a method for correcting the baseline (BL) in electrocardiographic (ECG) signals, the merit of the proposed methodology is based on dynamic programming (DP) and local sequence alignment. The fundamental objective is to achieve an effective reconstruction of the baseline of the signal, in order to be able to subtract the baseline signal from the original signal. By aligning the sequences, the idea is to be able to create a segmenting of all the waves contained in the signal (PQRSTwaves). Once the waves from the ECG have been aligned and segmented, a nodes localization is carried out which would represent an isoelectric line of the electrocardiographic signal, that is, nodes that fall between the PR and ST segments must be searched and accounted for. After identifying the data point that represent the isoelectric line of the signal, these are approximated by an interpolating function to achieve an approximation of what would correspond to the baseline of the ECG signal. The reconstructed signal is then removed from the original signal in order to achieve an adequate correction of the data. With the goal of obtaining an evaluation for the proposed methodology, the results are compared with methods where the use of wavelet transform was employed for the removal of ECG noises in the baseline signals.