Ultrasound Image Enhancement via Deconvolution

Abstract

Ultrasound imaging is a widespread medical imaging technique that is still a subject of active research. An important problem in ultrasound imaging is related to the effective use of deconvolution to improve the image quality. One of the ways to solve this problem is based on Robust Principal Component Analysis (RPCA), which requires some knowledge of the point spread function (PSF). Blind deconvolution (BD) refers to the situation when there is no prior knowledge of the PSF is available, and therefore, the latter needs to be estimated based on the given radio-frequency (RF) data. This report investigates the problem of blind deconvolution where the PSF being estimated can either be stationary or non-stationary. To evaluate the strengths and weaknesses of different deconvolution methods under consideration, we utilize two datasets corresponding to the PICMUS simulated phantom and the in vivo carotid artery, and apply multiple image quality indicators such as tissue-to-clutter ratio (TCR), contrast-to-noise ratio (CNR), full width at half maximum (FWHM), and others.