Skripsi
PENERAPAN KOMBINASI METODE EMPIRICAL MODE DECOMPOSITION DAN DISCRETE WAVELET TRANSFORM PADA PERBAIKAN BEAT SINYAL ELEKTROKARDIOGRAM
The electrocardiogram (ECG) is a medical examination method used to record the electrical activity of the heart and plays an important role in detecting cardiac abnormalities such as atrial fibrillation. An ECG signal is considered good when the P wave, QRS complex, and T wave patterns are clearly visible and free from noise contamination. Noise is an unwanted interference signal that contaminates ECG recordings due to body movements, muscle activity, or medical device errors during the recording process. The presence of noise causes wave patterns to become unclear and may affect the accuracy of clinical analysis. This study aims to perform ECG signal denoising using a combination of Empirical Mode Decomposition (EMD) and Discrete Wavelet Transform (DWT) methods for atrial fibrillation arrhythmia. The EMD method is applied in the initial stage to decompose the signal into several Intrinsic Mode Functions (IMFs), where IMFs identified as noise are removed while the main signal information is preserved. The DWT method is then applied after EMD to perform further filtering through Low Pass Filter (LPF) and High Pass Filter (HPF) processes to eliminate the remaining noise. This study produced an average Signal-to-Noise Ratio (SNR) value of 32.13 dB, a Root Mean Square Error (RMSE) value of 0.018 mV, and a Percent Root Difference (PRD) value of 4.3%. An SNR value of 32.13 dB indicates that the denoised ECG signal is much more dominant than the remaining noise, reflecting very good signal quality. An RMSE value of 0.018 mV indicates that the amplitude difference between the denoised signal and the original signal is very small, meaning that the waveform shape is well preserved. A PRD value of 4.3% indicates that the overall change in the signal shape is very small, so the characteristics of the P wave, QRS complex, and T wave are well maintained after the denoising process. These results demonstrate that the combination of EMD and DWT methods provides excellent performance in reducing noise while preserving the morphology of the ECG signal.
No other version available