Sudden Cardiac Death (SCD) is one of the leading causes of death in current society. Although increasing efforts have been done in clinical care, the survival rate of SCD remains low. To improve patient outcomes as well as actively prevent SCD, we aim at designing an energy-efficient intelligent system that can predict the cardiac health status and provide real-time warnings for users outside of hospitals. We achieve this by first developing features based on time-frequency scalogram extracted from ECG and then training a support vector machine (SVM) for classification of adverse pre-SCD events. The positive predictive value is reported to be over 95% on the standard MIT-BIH dataset. We also design and implement a versatile ECG processor with high functional flexibility and low power consumption to meet the requirements of an ideal real-time SCD prevention system.
This work was developed during my master study at Graduate Institute of Electronics Engineering, National Taiwan University (NTU), advised by Prof. Liang-Gee Chen. The thesis can be accessed through DOI:10.6342/NTU.2013.00511.