Three-dimensional time-domain full waveform inversion for sound speed and attenuation reconstruction in ultrasound computed tomography

Ultrasound computed tomography (USCT) is a noninvasive biomedical imaging modality that offers insights into acoustic properties such as the sound speed (SS) and acoustic attenuation (AA) of the human body, enhancing diagnostic accuracy and therapy planning. Full waveform inversion (FWI) is a promising USCT image reconstruction method that optimizes the parameter fields of a wave propagation model via gradient-based optimization. However, two-dimensional FWI methods are limited by their inability to account for three-dimensional wave propagation in the elevation direction, resulting in image artifacts. To address this problem, we propose a three-dimensional time-domain full waveform inversion algorithm to reconstruct the SS and AA distributions on the basis of a fractional Laplacian wave equation, adjoint field formulation, and gradient descent optimization. Validated by two sets of simulations, the proposed algorithm has potential for generating high-resolution and quantitative SS and AA distributions. This approach holds promise for clinical USCT applications, assisting early disease detection, precise abnormality localization, and optimized treatment planning, thus contributing to better healthcare outcomes.