Navegar por los elementos (1 total)
- Resumen es exacto "Cardiac ischemia is the leading cause of death in the world, as a consequence it is important to prevent and detect these events earlier. Traditionally, ischemia is detected by analyzing the level of the ST segment on the electrocardiogram (ECG). Although numerous methods have been used to improve the detection of ischemic events, interpretation of the ECG by a professional is still necessary in health centers. This is because the detection algorithms were intentionally created to obtain high sensitivity at the expense of specificity, thus detecting a large number of false positives. The main source of false positives are events that produce alterations in the ST level but are not caused by an ischemia, such as changes in heart rate, changes in ventricular conduction and in the direction of the cardiac electrical axis. In this work, with the aim of reducing the high rate of false positives, a classifier of ischemic and non-ischemic events was designed, where two new parameters obtained from applying the Continuous Wavelet Transform (CWT) to the ECG signal, were added to the parameters commonly used in the temporal domain. For this, the Long Term ST Database was used, a database specially designed to study this problem. It contains Holter records of approximately 24 hours with annotations made by specialists indicating the beginning and end of ischemic events and non-ischemic events, commonly interpreted by detectors as ischemic events. The performance obtained by our classifier in this database was a sensitivity and specificity of 84% and 93%, respectively. With the results obtained, it was concluded that the parameters proposed from the CWT were useful to successfully distinguish ischemic from nonischemic events. Based on the encouraging results of the first work, it was decided to analyze the ECG signal using Empirical Mode Decomposition (EMD) to perform an ischemia detector. The advantage of this decomposition is that it can be used for nonlinear and non-stationary signals. In this way, the ECG signal was decomposed into its intrinsic mode functions from which two parameters were extracted. Based on these parameters, an ischemia detector was designed. To evaluate the performance of the detector, the records corresponding to the MLIII and V4 leads of the European ST-T Database were used. This database was created specifically to evaluate ischemia detectors, since all registries have ischemic events annotated by specialists. The sensitivity and positive predictivity obtained when evaluating the detector in registers of MLIII lead was of 88%. When using the multi-lead detection with leads MLIII and V4, the sensitivity and positive predictivity obtained was of 92% and 80%, respectively. The results obtained demonstrated the usefulness of this decomposition and of the extracted parameters for the detection of cardiac ischemia. With the results of both works, it was possible to conclude that the techniques used, Continuous Wavelet Transformation and Empirical Mode Decomposition, provided relevant information to improve the detection and classification of ischemic events.
"
Ordenar por: