With cardiovascular disease (CVD) being one of the top causes of death in Canada, a solution must be developed to help treat patients in the early stages of CVD. In addition, with increasing patient waiting times, the demand for a system that can diagnose heart disease is also increasing. The ECG Analyzer provides a method of monitoring the heart and diagnosing heart disease in real-time using pattern recognition. The analyzer incorporates a feature extraction component and a classification component. Feature extraction uses QRS detection to acquire disease characterizing information for subsequent use in the classification component. The classification component is achieved through the use of support vector machines (SVM). Results demonstrated that the QRS detection algorithm and SVM classification performed reasonably well with classification error rates as low as 19.09%. Different kernel functions were used and the polynomial function was found to be the best option. At the conclusion of testing, it was noted that classification accuracy could be increased by using a higher dimensional feature vector in conjunction with feature selective classification.
Tisma, Robert, "Design of Software for an Electrocardiogram Analyzer" (2010). EE 4BI6 Electrical Engineering Biomedical Capstones. Paper 32.