Utilization of singular value decomposition in high-frame-rate cardiac blood flow imaging

Visualization of blood flow is important to evaluate the cardiac function. In this study, we propose two signal processing techniques based on singular value decomposition (SVD) for the visualization of blood flow: one is a filtering for clutter reduction and the other is a regularization method for color Doppler images. In the clutter filtering, contrast-to-noise ratio obtained by the SVD filtering was better (15.7 dB) than that by conventional finite impulse response filtering (-0.5 dB). In color flow imaging, the standard deviation of cardiac blood flow was decreased from 0.022 to 0.014 mm s^-1 by the proposed regularization method.