Performance study of wavelet-based ECG analysis for ST-segment detection

Heart disease has the second highest mortality rate in Japan. In particular, the ischemic heart disease such as the myocardial infarction and the heart failure requires emergency treatment. Although their symptoms emerge as an elevation or depression of the ST segment in ECG (electrocardiogram) waveform, it is very difficult to detect it because the S and T waves are quite small as compared with the R-wave. In this paper, we evaluate the practical utility of wavelet-based ECG analysis algorithm through the actual patient data. We used an algorithm proposed in literature. We reveal that the algorithm performs accurate waveform detection of ST waves in most cases but it fails when the ECG waveform is influenced by a large baseline fluctuation or by strong ST elevations that deform the ECG waveform significantly. In addition, we analyze the reasons why the algorithm fails to detect the ST waves.