Deep Learning for Ultrasound Image Formation: CUBDL Evaluation Framework and Open Datasets

被引：63

作者：

Hyun, Dongwoon ^{[1
]}

Wiacek, Alycen ^{[2
]}

Goudarzi, Sobhan ^{[3
]}

Rothlubbers, Sven ^{[4
]}

Asif, Amir ^{[5
]}

Eickel, Klaus ^{[6
]}

Eldar, Yonina C. ^{[7
]}

Huang, Jiaqi ^{[8
]}

Mischi, Massimo ^{[9
]}

Rivaz, Hassan

Sinden, David ^{[4
]}

van Sloun, Ruud J. G. ^{[9
,10
]}

Strohm, Hannah ^{[4
]}

Bell, Muyinatu A. Lediju ^{[11
,12
]}

机构：

[1] Stanford Univ, Dept Radiol, Stanford, CA 94305 USA

[2] Johns Hopkins Univ, Dept Elect & Comp Engn, Baltimore, MD 21218 USA

[3] Concordia Univ, Dept Elect & Comp Engn, Montreal, PQ H3G 1M8, Canada

[4] Fraunhofer Inst Digital Med MEVIS, D-28359 Bremen, Germany

[5] York Univ, Dept Elect Engn & Comp Sci, N York, ON M3J 1P3, Canada

[6] Univ Bremen, Dept Phys & Elect Engn, D-28359 Bremen, Germany

[7] Weizmann Inst Sci, Dept Math & Comp Sci, IL-7610001 Rehovot, Israel

[8] Johns Hopkins Univ, Dept Biomed Engn, Baltimore, MD 21218 USA

[9] Eindhoven Univ Technol, Dept Elect Engn, NL-5612 Eindhoven, Netherlands

[10] Philips Res, NL-5656 Eindhoven, Netherlands

[11] Johns Hopkins Univ, Dept Biomed Engn, Dept Elect & Comp Engn, Baltimore, MD 21218 USA

[12] Johns Hopkins Univ, Dept Comp Sci, Baltimore, MD 21218 USA

来源：

IEEE TRANSACTIONS ON ULTRASONICS FERROELECTRICS AND FREQUENCY CONTROL | 2021年 / 68卷 / 12期

基金：

英国工程与自然科学研究理事会; 美国国家卫生研究院; 加拿大自然科学与工程研究理事会;

关键词：

Ultrasonic imaging; Task analysis; Deep learning; Phantoms; In vivo; Array signal processing; Training; Beamforming; channel data; deep learning benchmark; evaluation metrics; neural networks; open science; sound speed estimation; ultrasound image formation; PHASE-ABERRATION CORRECTION; SPEED; NETWORK; IMPACT; INPUT;

D O I：

10.1109/TUFFC.2021.3094849

中图分类号：

O42 [声学];

学科分类号：

070206 ; 082403 ;

摘要：

Deep learning for ultrasound image formation is rapidly garnering research support and attention, quickly rising as the latest frontier in ultrasound image formation, with much promise to balance both image quality and display speed. Despite this promise, one challenge with identifying optimal solutions is the absence of unified evaluation methods and datasets that are not specific to a single research group. This article introduces the largest known international database of ultrasound channel data and describes the associated evaluation methods that were initially developed for the challenge on ultrasound beamforming with deep learning (CUBDL), which was offered as a component of the 2020 IEEE International Ultrasonics Symposium. We summarize the challenge results and present qualitative and quantitative assessments using both the initially closed CUBDL evaluation test dataset (which was crowd-sourced from multiple groups around the world) and additional in vivo breast ultrasound data contributed after the challenge was completed. As an example quantitative assessment, single plane wave images from the CUBDL Task 1 dataset produced a mean generalized contrast-to-noise ratio (gCNR) of 0.67 and a mean lateral resolution of 0.42 mm when formed with delay-and-sum beamforming, compared with a mean gCNR as high as 0.81 and a mean lateral resolution as low as 0.32 mm when formed with networks submitted by the challenge winners. We also describe contributed CUBDL data that may be used for training of future networks. The compiled database includes a total of 576 image acquisition sequences. We additionally introduce a neural-network-based global sound speed estimator implementation that was necessary to fairly evaluate the results obtained with this international database. The integration of CUBDL evaluation methods, evaluation code, network weights from the challenge winners, and all datasets described herein are publicly available (visit https://cubdl.jhu.edu for details).

引用

页码：3466 / 3483

页数：18

共 61 条

[1] The direct estimation of sound speed using pulse-echo ultrasound [J].