Artificial intelligence

Abstract: The goal of that work is check of the effectiveness of the presented EMD-method and the Widrow-Hoff gradient LMS-method for the baseline wander removal at ICP and electrocardiogram (ECG) signals, and comparison of the suggested method with statistically direct algorithms. The removal of such interference is a very important step in the preprocessing stage of essential medical signals for getting desired signal for clinical diagnoses. At this article a new method signal filtering was presented, in which the reconstruction of the reference signal is conditioned by lower frequency IMFs. This method does not use any preprocessing and post processing, and does not require prior estimates. The proposed filtering scheme, as compared to the widely used of a two-stage moving-average filter, lowpass-IIR and median filters, showed the effective baseline wander removal of ICP and EKG of signals without distortion of their waveform signals.

Keywords: baseline wander, empirical mode decomposition, intrinsic mode functions, adaptive filter, intracranial pressure, electrocardiogram.

References: