Ecg arrhythmia recognition improvement using respiration information

Abstract

Electrocardiogram (ECG) and respiration signals are two basic but important biomedical signals. They provide good source of information used to determine the patient's conditions, where the earlier is more popular. The difficulty is the ECG signals are usually of small amplitude and are susceptible to various noises such as: the 50 Hz grid noise, poor electrodes’ contacts with the patient's skin, the patient's emotional variations, the respiration and movements (including the breathing movements) of the patient, etc. In this paper we propose two ways to improve the accuracy of ECG signal recognition by filtering out the effect of the respiration in the ECG signal and by using the information of breathing stage as features in ECG signal classification. These approaches can improve the reliability and accuracy of the arrhythmia classification. As the classifier we use the modified neuro-fuzzy TSK network. The proposed solution will be tested with data from the MIT-BIH and the MGH/MF databases.