Image-Based Reconstruction of Tissue Scatterers Using Beam Steering for Ultrasound Simulation

Abstract

Numerical simulation of ultrasound images can facilitate the training of sonographers. An efficient and realistic model for the simulation of ultrasonic speckle is the convolution of the ultrasound point-spread function with a distribution of point scatterers. Nevertheless, for a given arbitrary tissue type, a scatterer map that would generate a realistic appearance of that tissue is not known a priori. In this paper, we introduce a principled approach to estimate (reconstruct) such a scatterer map from images, by solving the inverse-problem of ultrasound speckle formation, such that images from arbitrary view angles and transducer settings can be generated from those scatterer maps later in simulations. Robust reconstructions are achieved by using multiple measurements of the same tissue with different viewing parameters. For this purpose, a novel use of beam-steering to rapidly and conveniently acquire multiple images of the same scene is proposed. We demonstrate in numerical and physical phantoms and images that the appearance of synthesized images closely match real images for a range of viewing parameters and probe settings. We also present a scene editing scenario exploiting these scatterer representations to create realistic images of augmented anatomy.

Scatgan for Reconstruction of Ultrasound Scatterers Using Generative Adversarial Networks

“…Our method has shown to be relatively robust to changes in viewing parameters and US probe settings such as transducer rotation relative to the tissue region of interest. It performs well on synthetic data, and generalizes to in vivo data with relative ease, while remaining orders of magnitude faster at inference time than the state-of-the-art optimization-based alternative [6]. Our fast runtime would make an extension to 3D feasible, which will then enable the determination of realistic input scatter maps for ray-based training simulations [3].…”

“…From this scatterer map, a Comparison with Related Work. We quantitatively compare our results with that of [6], which performed a similar set of experiments. Their optimization-based technique is called "ScatRec1" when reconstructing scatter maps from a single B-mode image, and "ScatRec7±30 • " when reconstructing them from 7 images taken at regular beam-steering angles {−30 • , −25 • , ..., 30 • }.…”

“…Inverse Problem of Scatterer Reconstruction. Finding a scatter map g given an RF image can be seen in discrete space as the inverse problem of deconvolution [5,6], i.e. Hg + Γ=f , where H is the convolution matrix induced by the PSF, g is a vector of scatterer amplitudes, f is the resulting RF image, and Γ is the noise vector.…”

“…a solution g may not faithfully represent the tissue under different imaging settings, such as viewing directions. Mattausch et al [6] proposed multiple acquisitions for the same field-of-view via electronic beam steering, to obtain multiple equations:…”

“…Comparison of ScatGAN with iterative optimization based deconvolution method (ScatRec1,7) from[6] for inclusion (in silico) dataset, Experiment 2, 0-30 • .…”