Detection of ECG Fiducial Points Using Recursive Estimation and Kalman filtering

Abstract

QRS complex detection is regarded as a baseline procedure for the segmentation of electrocardiographic (ECG) signals, as it is usually the most distinctive component of the signal. Unfortunately, many QRS detection algorithms do not work well in pathological heartbeats, where QRS morphology changes radically. This paper addresses QRS detection by using a novel approach based on recursive estimation of the QRS envelope using Kalman Filter and smoothness priors. This approach effectively estimates fiducial points, as it considers an interval-dependent adaptive threshold, which is independent of the heartbeat morphology, reaching a robust detection. In order to validate this proposal, the MIT-BH, QT, and ST-T databases were used. A global accuracy of 99.4% with a sensitivity of 96.9% was achieved. The experimental results demonstrated an improvement of the proposed Kalman filter, showing that the performance is stable, maintaining a high performance as the noise level increases.

EEG Eye Blink Artifacts Removal with Wavelet Denoising and Bandpass Filtering

“…The underlying signal of the neural source is released to the processed FP1 signal. The Kalman filter is for removing a heavy artifact form of artifact 44 . It has been effective in identifying a second window containing various kinds of objects such as muscles and motion.…”