Tomographic Estimation of the Point-Spread Function of an Ultrasound Imaging System

The usefulness of ultrasound images in medical diagnostics is limited by inherent blurring and consequent poor image resolution. Algorithms that attempt to correct this blur generally require knowledge of the imaging system's point-spread function (PSF). In this report, we introduce a method for determining the three-dimensional PSF of an ultrasound imaging system by acquiring images of a line target and applying tomographic reconstruction methods. We show that imaging a line target yields the projection of the PSF along the axis of the target, and given a sucient number of projections at dierent angles, we can reconstruct the PSF tomographically, e.g. by using the ltered backprojection method. We discuss some of the practical considerations that need to be addressed when preparing a suitable line target phantom. Our experimental results show that the PSF estimated in this way has good delity to its theoretically predicted counterpart. We also include results of frequency-domain deconvolution with Tikhonov regularisation which show modest improvement in resolution as measured by the widths of the corresponding two-dimensional autocorrelation functions.