| Summary: | 碩士 === 國立雲林科技大學 === 資訊工程系 === 105 === In recent years, cardiovascular disease is the major causes of death. Cardiovascu-lar disease is mainly observed through electrocardiogram. ECG is susceptible to inter-ference from external factors, leading to added noise of ECG signal. In this study, ECG feature extraction combined with blood pressure.
In this study, the original signal uses median filtering to remove low-frequency noise, and then use Wavelet modulus maxima detection characteristics. The pulse transit time(PTT) is used to train the classifier used to identify the characteristics of the ECG.
In conclusion, the ECG signals add the noise for verifying the accuracy of our al-gorithm. The accuracy of our algorithm reached 95.76% under the strong noise signals.
|
|---|
