Full-wave attenuation reconstruction in the time domain for ultrasound computed tomography

Abstract

Acoustical attenuation (AA) maps in Ultrasound Computed Tomography (USCT) provide enhanced contrast between tissues compared to the speed of sound (SS), which is the most common property of tissue studied with this technique. Currently, the full wave inversion (FWI) methods used for their reconstruction are very different: the AA is mainly estimated using frequency domain algorithms, while the SS is more often recovered in the time domain. In this work we present a novel strategy to recover the attenuation maps through a straightforward and simplified procedure in the time domain. A gradient descent method was employed to optimize iteratively the attenuation distribution. The expression for the functional gradient of the norm of the global deviation between experimental and simulated data was obtained using an adjoint method. The optimization code, implemented in C++, employs a CUDA version of the k-Wave software to perform forward and backward wave propagation. Noisy simulated data was used to test the performance of the proposed method. The simplicity of the formulation of this new method may facilitate the reconstruction of AA and SS maps under a common framework in USCT.