Determination of Morphologically Characteristic PCG Segments from Spectrogram Image.docVIP

74 Telfor Journal, Vol. 2, No. 2, 2010. Determination of Morphologically Characteristic PCG Segments from Spectrogram Image* Ana M. Gavrovska, Student Member, IEEE, Milorad P. Paska?, Student Member, IEEE and Irini S. Reljin, Senior Member, IEEE necessary to characterize appropriate events. Joint time- Abstract phonocardiac amplitude and frequency, allows the determination of morphological characteristic segments in phonocardiogram (PCG), both in short and long sequences. For this purpose, — The signal, three-dimensional simultaneously presentation of frequency analysis allows the simultaneous examination of both domains and as result we get a three-dimensional presentation of PCG. Simultaneous time, amplitude and frequency presentation represents an excellent way for considering time, the STFT (Short-Time Fourier Transform) spectrogram precise determination of cardiac events and noise origin. images were used. By applying some methods known from image processing it is possible to recognize and extract basic heart sounds and murmurs from such time-frequency images. The method is tested over several characteristic test phonocardiac signals. The purpose of this paper is to introduce morphological processing of PCG spectrogram image as a tool for distinguishing basic heart sounds and murmurs. This approach was applied to STFT (Short-Time Fourier Transform) images that are briefly discussed in section II. In section III we give a short overview of the basic concepts of morphological processing. The problem that we want to solve here is defined in section IV. Method of analysis is described in section V, where we conducted testing on five different characteristic signals with confirmed diagnosis. Section VI concludes the paper and it is mainly devoted to future research. Keywords — Morphological process