Derivation of Respiratory Signal from Single- Channel ECGs Based on Source Statistics

Respiration derivation from ECG signals would be an alternative approach for obtaining respiration related knowledge that is especially useful in the situation in which the respiration is not routinely monitored by specialized equipment. Although several published methods could give reasonable results for this purpose, for some applications, they still have drawbacks in some respects, e.g., requiring multi-channel ECG signal, being robust-less against noise, etc.. In order to overcome these drawbacks, in this paper, we investigate the possibility of obtaining respiratory knowledge from a single-channel ECG signal based on its higher order statistics, which is a notable feature that the conventional methods do not hold. Another worthy of mention feature is that the method depends on Rwave detections only, such that other feature detections, e.g., Q-wave or S-wave detection, are not necessary anymore for the processing. This feature does not make the processing simpler only, with the feature related to higher order statistics together, but it also strengthens the robustness against noise. Our experiments show the improved performances with the method.