Background: Fast and accurate quality estimation of the electrocardiogram (ECG) signal\nis a relevant research topic that has attracted considerable interest in the scientific\ncommunity, particularly due to its impact on tele-medicine monitoring systems, where\nthe ECG is collected by untrained technicians. In recent years, a number of studies have\naddressed this topic, showing poor performance in discriminating between clinically\nacceptable and unacceptable ECG records.\nMethods: This paper presents a novel, simple and accurate algorithm to estimate the\nquality of the 12-lead ECG by exploiting the structure of the cross-covariance matrix\namong different leads. Ideally, ECG signals from different leads should be highly correlated\nsince they capture the same electrical activation process of the heart. However, in\nthe presence of noise or artifacts the covariance among these signals will be affected.\nEigenvalues of the ECG signals covariance matrix are fed into three different supervised\nbinary classifiers.\nResults and conclusion: The performance of these classifiers were evaluated using\nPhysioNet/CinC Challenge 2011 data. Our best quality classifier achieved an accuracy\nof 0.898 in the test set, while having a complexity well below the results of contestants\nwho participated in the Challenge, thus making it suitable for implementation in current\ncellular devices.
Loading....