Doppler Signal Processing of Blood Flow Using a Wavelet Transform

Abstract

In medical echography, Doppler signal analysis is one of the most important diagnostic techniques. In Doppler signal analysis, time-frequency localization techniques are necessary. Many techniques have been proposed for the analysis. Short-time Fourier transform (STFT) is one of the most popular techniques. STFT has a fixed time-frequency window making it inappropriate for analysis of signals having a wide bandwidth. On the other hand, recently, wavelet transform (WT) has received much attention. WT has a flexible time-frequency window adaptable to any time-frequency range. In this paper, we describe an application of STFT and WT to Doppler signal processing. The STFT in this study had a rectangular window and required fast Fourier transform (FFT). On the other hand, the WT with the Morlet function as the mother wavelet was modified to reduce the amount of calculation. Then, we compared the suitability of STFT and WT for application to Doppler signal processing of a simulation and an actual blood flow. In the simulation, a flow having fast and slow velocity changes was assumed. The simulation revealed that the WT could reproduce the fast change of flow velocity, while the STFT could not. Then, a blood flow Doppler signal was measured at a point near the mitral valve in the left ventriculus. Only the WT was capable of reproducing the fast change of flow velocity in the heart. The WT is expected to be useful for analyzing blood flow disorders such as regurgitation.