ENAS-B: Combining ENAS With Bayesian Optimization for Automatic Design of Optimal CNN Architectures for Breast Lesion Classification From Ultrasound Images.

IF 2.5 4区 医学 Q1 ACOUSTICS Ultrasonic Imaging Pub Date : 2024-01-01 Epub Date: 2023-11-20 DOI:10.1177/01617346231208709
Mohammed Ahmed, Hongbo Du, Alaa AlZoubi
{"title":"ENAS-B: Combining ENAS With Bayesian Optimization for Automatic Design of Optimal CNN Architectures for Breast Lesion Classification From Ultrasound Images.","authors":"Mohammed Ahmed, Hongbo Du, Alaa AlZoubi","doi":"10.1177/01617346231208709","DOIUrl":null,"url":null,"abstract":"<p><p>Efficient Neural Architecture Search (ENAS) is a recent development in searching for optimal cell structures for Convolutional Neural Network (CNN) design. It has been successfully used in various applications including ultrasound image classification for breast lesions. However, the existing ENAS approach only optimizes cell structures rather than the whole CNN architecture nor its trainable hyperparameters. This paper presents a novel framework for automatic design of CNN architectures by combining strengths of ENAS and Bayesian Optimization in two-folds. Firstly, we use ENAS to search for optimal normal and reduction cells. Secondly, with the optimal cells and a suitable hyperparameter search space, we adopt Bayesian Optimization to find the optimal depth of the network and optimal configuration of the trainable hyperparameters. To test the validity of the proposed framework, a dataset of 1522 breast lesion ultrasound images is used for the searching and modeling. We then evaluate the robustness of the proposed approach by testing the optimized CNN model on three external datasets consisting of 727 benign and 506 malignant lesion images. We further compare the CNN model with the default ENAS-based CNN model, and then with CNN models based on the state-of-the-art architectures. The results (error rate of no more than 20.6% on internal tests and 17.3% on average of external tests) show that the proposed framework generates robust and light CNN models.</p>","PeriodicalId":49401,"journal":{"name":"Ultrasonic Imaging","volume":" ","pages":"17-28"},"PeriodicalIF":2.5000,"publicationDate":"2024-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Ultrasonic Imaging","FirstCategoryId":"5","ListUrlMain":"https://doi.org/10.1177/01617346231208709","RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"2023/11/20 0:00:00","PubModel":"Epub","JCR":"Q1","JCRName":"ACOUSTICS","Score":null,"Total":0}
引用次数: 0

Abstract

Efficient Neural Architecture Search (ENAS) is a recent development in searching for optimal cell structures for Convolutional Neural Network (CNN) design. It has been successfully used in various applications including ultrasound image classification for breast lesions. However, the existing ENAS approach only optimizes cell structures rather than the whole CNN architecture nor its trainable hyperparameters. This paper presents a novel framework for automatic design of CNN architectures by combining strengths of ENAS and Bayesian Optimization in two-folds. Firstly, we use ENAS to search for optimal normal and reduction cells. Secondly, with the optimal cells and a suitable hyperparameter search space, we adopt Bayesian Optimization to find the optimal depth of the network and optimal configuration of the trainable hyperparameters. To test the validity of the proposed framework, a dataset of 1522 breast lesion ultrasound images is used for the searching and modeling. We then evaluate the robustness of the proposed approach by testing the optimized CNN model on three external datasets consisting of 727 benign and 506 malignant lesion images. We further compare the CNN model with the default ENAS-based CNN model, and then with CNN models based on the state-of-the-art architectures. The results (error rate of no more than 20.6% on internal tests and 17.3% on average of external tests) show that the proposed framework generates robust and light CNN models.

查看原文
分享 分享
微信好友 朋友圈 QQ好友 复制链接
本刊更多论文
ENAS- b:结合ENAS和贝叶斯优化自动设计乳腺病变超声图像分类的最优CNN架构。
高效神经结构搜索(ENAS)是卷积神经网络(CNN)设计中搜索最优细胞结构的最新发展。它已成功地应用于各种应用,包括乳腺病变的超声图像分类。然而,现有的ENAS方法只优化单元结构,而不是整个CNN架构或其可训练的超参数。本文结合ENAS和贝叶斯优化的优点,提出了一种新的CNN结构自动设计框架。首先,我们使用ENAS搜索最优的正常细胞和还原细胞。其次,利用最优单元和合适的超参数搜索空间,采用贝叶斯优化方法求出网络的最优深度和可训练超参数的最优配置;为了验证所提框架的有效性,使用1522张乳腺病变超声图像数据集进行搜索和建模。然后,我们通过在由727张良性和506张恶性病变图像组成的三个外部数据集上测试优化后的CNN模型来评估所提出方法的鲁棒性。我们进一步将CNN模型与默认的基于enas的CNN模型进行比较,然后与基于最先进架构的CNN模型进行比较。结果(内部测试错误率不超过20.6%,外部测试错误率平均不超过17.3%)表明,所提框架生成的CNN模型鲁棒轻巧。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 去求助
来源期刊
Ultrasonic Imaging
Ultrasonic Imaging 医学-工程:生物医学
CiteScore
5.10
自引率
8.70%
发文量
15
审稿时长
>12 weeks
期刊介绍: Ultrasonic Imaging provides rapid publication for original and exceptional papers concerned with the development and application of ultrasonic-imaging technology. Ultrasonic Imaging publishes articles in the following areas: theoretical and experimental aspects of advanced methods and instrumentation for imaging
期刊最新文献
Image Features and Diagnostic Value of Contrast-Enhanced Ultrasound for Ductal Carcinoma In Situ of the Breast: Preliminary Findings. Ultrasonic Imaging of Deeper Bone Defect Using Virtual Source Synthetic Aperture with Phased Shift Migration: A Phantom Study. Predictive Value of the Nomogram Model Based on Multimodal Ultrasound Features for Benign and Malignant Thyroid Nodules of C-TIRADS Category 4. High Frequency Ultrasound Transducer Based on Sm-Doped Pb(Mg1/3Nb2/3)O3-0.28PbTiO3 Ceramic for Intravascular Ultrasound Imaging. Development and Assessment of a Predictive Model for Ki-67 Expression Using Ultrasound Indicators and Non-Morphological Magnetic Resonance Imaging Parameters Before Breast Cancer Therapy.
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
现在去查看 取消
×
提示
确定
0
微信
客服QQ
Book学术公众号 扫码关注我们
反馈
×
意见反馈
请填写您的意见或建议
请填写您的手机或邮箱
已复制链接
已复制链接
快去分享给好友吧!
我知道了
×
扫码分享
扫码分享
Book学术官方微信
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术
文献互助 智能选刊 最新文献 互助须知 联系我们:info@booksci.cn
Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。
Copyright © 2023 Book学术 All rights reserved.
ghs 京公网安备 11010802042870号 京ICP备2023020795号-1