Analisis Non-Stasioner pada Deteksi Non-Invasive Sinyal Suara Jantung Koroner

Abstract

Feature extraction has become a very important factor in electronic heart sound diagnosis system development. Therefore, it is necessary to conduct a research to find an effective feature extraction method. This research has learned more about feature extraction method, using non-stationary signal processing, Short Time Fourier Transform (STFT) to extract three coronary heart disease signals. The results show that the first signal has an average frequency of 505,56±8,82 Hz. This signal is detected on an average window of 21,44±2,92, and has an average time of 0,05±0,02 s. The second signal has an average frequency of 376,11±2,20 Hz, with average window of 141,67±2,5, and average time of 0,35±0,02 s. The results of feature extraction on the third signal shows an average frequency of 217,14±12,78 Hz, average window of 74,29±4,16, and the average time is 0,17±0,02 s. This results indicates that the entire frequency has an average value of more than 200 Hz.