Effect of filter selection on classification of extrasystole heart sounds via mobile devices

Sound of the heart is the basic biomedical information utilized for diagnosis by medical doctors in heart diseases. These sounds show differences according to different pathological characteristics. Extra systole sounds, which mean extra heartbeat, can be perceived as throbbing by people. Occurrence of these sounds in certain age groups may be the indication of tachycardia. In this study, effect of Butterworth, Chebyshev and Elliptic filters on classification results for noise removal in extra systole specific sounds in heart sound database is analyzed. The filters chosen and other methods are paid attention to be faster because the application developed for this aim will be used on mobile devices. Db5 type wavelet transformation method has been used to gain less as feature set. Support vector machine has been used to classify. According to the results gained, the fastest filter for noise removal in extra systole specific heart sounds is Butterworth and the filter that gives the best classification results is Elliptic filter.