Deep Learning for Accelerated Ultrasound Imaging

TL;DR

This paper introduces a deep learning method that reconstructs high-quality ultrasound images from limited data by interpolating missing RF data in the Fourier domain, enabling faster and more efficient imaging.

Contribution

It presents a novel deep learning approach that leverages Fourier domain sparsity to interpolate missing RF data, reducing data requirements without quality loss.

Findings

Effective reduction of data rate in ultrasound imaging

High-quality images reconstructed from sub-sampled RF data

Validated on real ultrasound system data

Abstract

In portable, 3-D, or ultra-fast ultrasound (US) imaging systems, there is an increasing demand to reconstruct high quality images from limited number of data. However, the existing solutions require either hardware changes or computationally expansive algorithms. To overcome these limitations, here we propose a novel deep learning approach that interpolates the missing RF data by utilizing the sparsity of the RF data in the Fourier domain. Extensive experimental results from sub-sampled RF data from a real US system confirmed that the proposed method can effectively reduce the data rate without sacrificing the image quality.