Phase Aberration Correction: A Convolutional Neural Network Approach

Abstract

One of the main sources of image degradation in ultrasound imaging is the phase aberration effect, which imposes limitations to both data acquisition and reconstruction. Phase aberration is induced by spatial changes in sound velocity compared to the default values and degrades the quality of beam focusing. In addition, it prevents received channel signals to be summed coherently. In this paper, for the first time, we propose a method to estimate the aberrator profile from an ultrasound B-mode image using a deep convolutional neural network (CNN) in order to compensate for the phase aberration effect. In contrast to traditional methods, which mostly apply time-consuming processing techniques on channel RF signals and need several iterations for a reasonable accuracy, the proposed approach is computationally efficient, and utilizes only the B-mode image to estimate the aberrator profile in one shot with a high accuracy. We experimentally investigate main characteristics of the proposed approach and present a quantitative evaluation of the estimated aberrator profile. The proposed method is compared with the conventional delay-and-sum (DAS) method and a method based on nearest-neighbor cross-correlation (NNCC). Results demonstrate that the proposed CNN method substantially outperforms other methods. INDEX TERMS Convolutional neural networks, deep learning, phase aberration, ultrasound.

Deep-learning-based skull-induced artifact reduction for transcranial ultrasound imaging: simulation study

“…Another feature of the proposed method is using channel data as both the input and output of the artifact reduction model since channel data carries more information about the imaging subject compared to b-mode data. Using channel data as the model input has been done in previous research [10]. However, a thorough comparison between channel data and bmode data as the model input to denoising or artifact reduction models has not been made yet.…”

Tunable Image Quality Control of 3-D Ultrasound using Switchable CycleGAN

“…Recently, various authors have proposed image enhancement algorithms and beamformers that use deep neural networks (Yoon et al (2019); Khan et al (2020a); Hyun et al (2019); Nair et al (2018); Brickson et al (2021); Solomon et al (2019); Sharifzadeh et al (2020); Vedula et al (2017); Perdios et al (2017Perdios et al ( , 2018; Van Sloun et al (2019); Luchies and Byram (2018); Khan et al (2019); Kokil and Sudharson (2020) 2021)). Unfortunately, this approach is not suitable for our 3-D US image enhancement problem as there are no matched high-resolution 3-D US images that can be used as ground truths for supervision.…”

Phase Aberration Robust Beamformer for Planewave US Using Self-Supervised Learning

“…Other deep CNN-based SoS calibration, estimation and reconstruction methods are presented in [26]- [28]. A deep CNN-based approach where the phase aberrator profiles are estimated using B-mode US images for phase aberration correction is presented in [29]. In particular, Sharifzadeh et al [29] designed a simulation study using convolution neural network.…”

“…A deep CNN-based approach where the phase aberrator profiles are estimated using B-mode US images for phase aberration correction is presented in [29]. In particular, Sharifzadeh et al [29] designed a simulation study using convolution neural network. Their CNN-based method was trained and tested on simulation phantom data to predict aberrator profile of individual elements.…”