ECG compressed sensing method with high compression ratio and dynamic model reconstruction

Abstract

This paper introduces an alternative method for compressed sensing and reconstruction of ECG that is patient agnostic and offers a high compression ratio. The high compression ratio is achieved by high decimation of the measurement signal and its post requantization, further decreasing the number of bits needed for information transfer. The sensing method also incorporates a QRS detector to detect exact R wave positions for signal segmentation before compression. ECG signal is also normalized in amplitude and offset, which maintains the bit resolution during requantization. The reconstruction employs a simple dynamic ECG model, parameters of which are calculated from the measurement signal by the Differential Evolution algorithm. The proposed method was evaluated using the MIT-BIH arrhythmia database and compared with two wavelet dictionary reconstruction methods. The proposed method keeps the structure of heartbeats preserved including the exact positions of R waves, and it reduces the noise interfering with ECG signals.