Article(id=1156907874640286636, tenantId=1146029695717560320, journalId=1146123166801305609, issueId=1156907871645556837, articleNumber=null, orderNo=null, doi=10.12404/j.issn.1671-1815.2309339, pmid=null, cstr=null, oa=null, hot=null, price=null, onlineType=0, articleFormat=0, articleType=null, articleTypeStr=research-article, receivedDate=1701014400000, receivedDateStr=2023-11-27, revisedDate=1719158400000, revisedDateStr=2024-06-24, acceptedDate=null, acceptedDateStr=null, onlineDate=1753757931622, onlineDateStr=2025-07-29, pubDate=1737993600000, pubDateStr=2025-01-28, doiRegisterDate=null, doiRegisterDateStr=null, onlineIssueDate=1753757931622, onlineIssueDateStr=2025-07-29, onlineJustAcceptDate=null, onlineJustAcceptDateStr=null, onlineFirstDate=null, onlineFirstDateStr=null, sourceXml=null, magXml=null, createTime=1753757931622, creator=13701087609, updateTime=1753757931622, updator=13701087609, issue=Issue{id=1156907871645556837, tenantId=1146029695717560320, journalId=1146123166801305609, year='2025', volume='25', issue='3', pageStart='879', pageEnd='1312', issueExtLink='null', onlineDate='null', pubDate='null', beforeIssueId=null, nextIssueId=null, price=null, status=1, issueComplete=1, articleOrder=1, issueType=-1, specialIssue=0, createTime=1753757930909, creator=13701087609, updateTime=1765095544280, updator=13701087609, preIssue=null, nextIssue=null, ext={EN=IssueExt(id=1204461268821320541, tenantId=1146029695717560320, journalId=1146123166801305609, issueId=1156907871645556837, language=EN, specialIssueTitle=, coverIllustrator=, specialIssueEditor=, specialIssueAbout=), CN=IssueExt(id=1204461268825514846, tenantId=1146029695717560320, journalId=1146123166801305609, issueId=1156907871645556837, language=CN, specialIssueTitle=, coverIllustrator=, specialIssueEditor=, specialIssueAbout=)}, issueFiles=null}, startPage=962, endPage=968, ext={EN=ArticleExt(id=1156907875697251254, articleId=1156907874640286636, tenantId=1146029695717560320, journalId=1146123166801305609, language=EN, title=CADCNet: An Improved Algorithm for Retinal Vessel Segmentation, columnId=1156262732384031076, journalTitle=Science Technology and Engineering, columnName=Papers·Medicine, runingTitle=null, highlight=null, articleAbstract=

Traditional retinal vessel segmentation methods often face challenges such as missegmentation caused by optic disc confusion, lack of continuity in segmentation results, and imprecise segmentation in detailed regions. To address these issues, a retinal vessel segmentation algorithm was proposed based on UNet. The algorithm replaced traditional square convolutions with a fusion of horizontal and vertical one-dimensional convolutions and two-dimensional square convolutions, enhancing the representation capability of the eye region. A multi-scale branch approach was adopted to increase feature space diversity, thereby improving the network’s feature learning and expression capabilities. Additionally, to further enhance segmentation performance, multi-layer dilated convolutions was introduced into the deep structure of the autoencoder, replacing traditional simple pooling operations. This approach enlarged the convolution kernel size and expanded the receptive field, achieving a fusion of multi-scale shallow and deep feature information. The proposed algorithm was evaluated on the public DRIVE and CHASE_DB1 datasets. Experimental results demonstrates that the algorithm achieves precision (0.956 8 and 0.959 8) and F1 scores (0.832 6 and 0.830 4), respectively. Compared with traditional UNet and recent UNet-based retinal vessel segmentation methods, the proposed algorithm shows advantages in accuracy, sensitivity, specificity, and F1 metrics, these validation results fully demonstrate the proposed model’s strong capability in precise segmentation tasks.

, correspAuthors=Ying-mei WANG, authorNote=null, correspAuthorsNote=null, copyrightStatement=null, copyrightOwner=null, extLink=null, articleAbsUrl=null, sourceXml=null, magXml=null, pdfUrl=null, pdf=null, pdfFileSize=null, pdfExtLink=null, richHtmlUrl=null, mobilePdfUrl=null, reviewReport=null, pdfFirstPage=null, abstractGraph=null, abstractGraphContent=null, abstractVideo=null, citation=null, cebUrl=null, magXmlContent=null, mapNumber=null, authorCompany=null, fund=null, authors=null, authorsList=Yu-chao YUE, Ying-mei WANG, Jia-chuan QIN), CN=ArticleExt(id=1156907917032116264, articleId=1156907874640286636, tenantId=1146029695717560320, journalId=1146123166801305609, language=CN, title=CADCNet: 一种改进的视网膜血管分割算法, columnId=1156262732526637414, journalTitle=科学技术与工程, columnName=论文·医药、卫生, runingTitle=null, highlight=null, articleAbstract=

目前传统的视网膜血管分割方法存在的视盘混淆引起的误分割、分割结果缺乏连续性,以及细节区域分割不精准等问题。为解决这一难题,提出了一种基于UNet的视网膜血管分割算法。该算法利用两个水平和垂直一维卷积和二维方形卷积的融合替代传统方形卷积,提高了眼球区域的表征能力;采用了多尺度分支增加特征空间的多样性,提升了网络的特征学习和表达能力。此外,为进一步改善分割效果,还将多层膨胀卷积引入自编码器的深层结构,替代了传统的简单池化操作,增大卷积核的大小,扩大了感受野范围,实现了多尺度浅层特征和深层特征信息的融合。本文算法在公开DRIVE和CHASE_DB1两个数据集上进行了评估,实验结果表明,本文算法的精确率和F1上分别达到了0.956 8、0.959 8和0.832 6、0.830 4。与传统的UNet和近期部分UNet改进网络视网膜血管分割方法相比,本文算法在准确率、敏感度、特异性、F1指标上表现出一定的优势,这一验证结果充分证明了本文所提出的模型在分割任务上具备较强的精确分割能力。

, correspAuthors=王迎美, authorNote=null, correspAuthorsNote=
* 王迎美(1987—),女,汉族,山东新泰人,博士,副教授。研究方向:医学图像处理与重建。E-mail:
, copyrightStatement=null, copyrightOwner=null, extLink=null, articleAbsUrl=null, sourceXml=hKBHt1v3e6sLJsmzBnOOQQ==, magXml=vaHR8Or2DuIYt5mX0CIcvg==, pdfUrl=null, pdf=VKMjQjxEx2H2IBqrUx7pyg==, pdfFileSize=8972085, pdfExtLink=null, richHtmlUrl=null, mobilePdfUrl=null, reviewReport=null, pdfFirstPage=null, abstractGraph=yafCbzbIr1fVTnGsMkkrEA==, abstractGraphContent=null, abstractVideo=null, citation=null, cebUrl=null, magXmlContent=A3aAwYgqmyjceB2gJQFJ5g==, mapNumber=null, authorCompany=null, fund=null, authors=

岳昱超(1998—),男,汉族,山东威海人,硕士研究生。研究方向:医学图像处理。E-mail:

, authorsList=岳昱超, 王迎美, 秦嘉川)}, authors=[Author(id=1204542855940383501, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, orderNo=0, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=yueyuchao1998@163.com, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1204542856032658197, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, authorId=1204542855940383501, language=EN, stringName=Yu-chao YUE, firstName=Yu-chao, middleName=null, lastName=YUE, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1. School of Mathematics and Statistics, Shandong University of Technology, Zibo 255000, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1204542856150098719, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, authorId=1204542855940383501, language=CN, stringName=岳昱超, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1.山东理工大学数学与统计学院, 淄博 255000, bio={"content":"

岳昱超(1998—),男,汉族,山东威海人,硕士研究生。研究方向:医学图像处理。E-mail:

"}, bioImg=null, bioContent=

岳昱超(1998—),男,汉族,山东威海人,硕士研究生。研究方向:医学图像处理。E-mail:

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1204542855709696750, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, xref=null, ext=[AuthorCompanyExt(id=1204542855718085360, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, companyId=1204542855709696750, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1. School of Mathematics and Statistics, Shandong University of Technology, Zibo 255000, China), AuthorCompanyExt(id=1204542855726473969, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, companyId=1204542855709696750, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1.山东理工大学数学与统计学院, 淄博 255000)])]), Author(id=1204542856250762028, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, orderNo=1, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=yingmeiwang@sdut.edu.cn, emailSecond=null, emailThird=null, correspondingAuthor=1, authorType=1, ext={EN=AuthorExt(id=1204542856351425332, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, authorId=1204542856250762028, language=EN, stringName=Ying-mei WANG, firstName=Ying-mei, middleName=null, lastName=WANG, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, *, address=1. School of Mathematics and Statistics, Shandong University of Technology, Zibo 255000, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1204542856494031680, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, authorId=1204542856250762028, language=CN, stringName=王迎美, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, *, address=1.山东理工大学数学与统计学院, 淄博 255000, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1204542855709696750, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, xref=null, ext=[AuthorCompanyExt(id=1204542855718085360, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, companyId=1204542855709696750, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1. School of Mathematics and Statistics, Shandong University of Technology, Zibo 255000, China), AuthorCompanyExt(id=1204542855726473969, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, companyId=1204542855709696750, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1.山东理工大学数学与统计学院, 淄博 255000)])]), Author(id=1204542856619860806, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, orderNo=2, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=null, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1204542856749884239, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, authorId=1204542856619860806, language=EN, stringName=Jia-chuan QIN, firstName=Jia-chuan, middleName=null, lastName=QIN, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=2, address=2. Shinva Medical Instrument Co., Ltd., Zibo 255086, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1204542856858936157, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, authorId=1204542856619860806, language=CN, stringName=秦嘉川, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=2, address=2.山东新华医疗器械股份有限公司, 淄博 255086, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1204542855822942973, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, xref=null, ext=[AuthorCompanyExt(id=1204542855831331581, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, companyId=1204542855822942973, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2. Shinva Medical Instrument Co., Ltd., Zibo 255086, China), AuthorCompanyExt(id=1204542855839720190, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, companyId=1204542855822942973, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2.山东新华医疗器械股份有限公司, 淄博 255086)])])], keywords=[Keyword(id=1204542858092061551, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, language=EN, orderNo=1, keyword=retinal vessel segmentation), Keyword(id=1204542858171753336, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, language=EN, orderNo=2, keyword=continuous dilation convolution), Keyword(id=1204542858280805248, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, language=EN, orderNo=3, keyword=deep learning), Keyword(id=1204542858389857162, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, language=EN, orderNo=4, keyword=asymmetric convolution), Keyword(id=1204542858477937556, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, language=EN, orderNo=5, keyword=UNet model), Keyword(id=1204542858553435036, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, language=CN, orderNo=1, keyword=视网膜血管分割), Keyword(id=1204542858700235685, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, language=CN, orderNo=2, keyword=连续膨胀卷积), Keyword(id=1204542858809287600, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, language=CN, orderNo=3, keyword=深度学习), Keyword(id=1204542858930922422, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, language=CN, orderNo=4, keyword=不对称卷积), Keyword(id=1204542859031585722, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, language=CN, orderNo=5, keyword=UNet模型)], refs=[Reference(id=1204542862781292618, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, doi=null, pmid=null, pmcid=null, year=2014, volume=2, issue=2, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[1], rfOrder=0, authorNames=Wong W L, Su X, Li X, journalName=The Lancet Global Health, refType=null, unstructuredReference=Wong W L, Su X, Li X, et al. Global prevalence of age-related macular degeneration and disease burden projection for 2020 and 2040: a systematic review and meta-analysis[J]. The Lancet Global Health, 2014, 2(2): e106-e116., articleTitle=Global prevalence of age-related macular degeneration and disease burden projection for 2020 and 2040: a systematic review and meta-analysis, refAbstract=null), Reference(id=1204542862902927444, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, doi=null, pmid=null, pmcid=null, year=2016, volume=123, issue=5, pageStart=1036, pageEnd=1042, url=null, language=null, rfNumber=[2], rfOrder=1, authorNames=Holden B A, Fricke T R, Wilson D A T R, journalName=Ophthalmology, refType=null, unstructuredReference=Holden B A, Fricke T R, Wilson D A T R, et al. Global prevalence of myopia and high myopia and temporal trends from 2000 through 2050[J]. Ophthalmology, 2016, 123(5): 1036-1042., articleTitle=Global prevalence of myopia and high myopia and temporal trends from 2000 through 2050, refAbstract=null), Reference(id=1204542863041339489, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, doi=null, pmid=null, pmcid=null, year=2014, volume=121, issue=11, pageStart=2081, pageEnd=2090, url=null, language=null, rfNumber=[3], rfOrder=2, authorNames=Tham Y C, Li X, Wong T Y, journalName=Ophthalmology, refType=null, unstructuredReference=Tham Y C, Li X, Wong T Y, et al. Global prevalence of glaucoma and projections of glaucoma burden through 2040: a systematic review and meta-analysis[J]. Ophthalmology, 2014, 121(11): 2081-2090., articleTitle=Global prevalence of glaucoma and projections of glaucoma burden through 2040: a systematic review and meta-analysis, refAbstract=null), Reference(id=1204542863158780008, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, doi=null, pmid=null, pmcid=null, year=2012, volume=96, issue=5, pageStart=614, pageEnd=618, url=null, language=null, rfNumber=[4], rfOrder=3, authorNames=Pascolini D, Mariotti S P, journalName=British Journal of Ophthalmology, refType=null, unstructuredReference=Pascolini D, Mariotti S P. Global estimates of visual impairment: 2010[J]. British Journal of Ophthalmology, 2012, 96(5): 614-618., articleTitle=Global estimates of visual impairment: 2010, refAbstract=null), Reference(id=1204542863251054702, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, doi=null, pmid=null, pmcid=null, year=2015, volume=null, issue=null, pageStart=3431, pageEnd=3440, url=null, language=null, rfNumber=[5], rfOrder=4, authorNames=Long J, Shelhamer E, Darrell T, journalName=Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, refType=null, unstructuredReference=Long J, Shelhamer E, Darrell T. Fully convolutional networks for semantic segmentation[C]// Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. New York: IEEE, 2015: 3431-3440., articleTitle=Fully convolutional networks for semantic segmentation, refAbstract=null), Reference(id=1204542863343329401, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, doi=null, pmid=null, pmcid=null, year=2015, volume=null, issue=null, pageStart=234, pageEnd=241, url=null, language=null, rfNumber=[6], rfOrder=5, authorNames=Ronneberger O, Fischer P, Brox T, journalName=Medical Image Computing and Computer-Assisted Intervention-MICCAI 2015: 18th International Conference, refType=null, unstructuredReference=Ronneberger O, Fischer P, Brox T. U-net: Convolutional networks for biomedical image segmentation[C]// Medical Image Computing and Computer-Assisted Intervention-MICCAI 2015: 18th International Conference. Berlin:Springer International Publishing, 2015: 234-241., articleTitle=U-net: Convolutional networks for biomedical image segmentation, refAbstract=null), Reference(id=1204542863456575617, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, doi=null, pmid=null, pmcid=null, year=2017, volume=null, issue=null, pageStart=472, pageEnd=480, url=null, language=null, rfNumber=[7], rfOrder=6, authorNames=Yu F, Koltun V, Funkhouser T, journalName=Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, refType=null, unstructuredReference=Yu F, Koltun V, Funkhouser T. Dilated residual networks[C]// Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. New York: IEEE, 2017: 472-480., articleTitle=Dilated residual networks, refAbstract=null), Reference(id=1204542863557238916, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, doi=null, pmid=null, pmcid=null, year=2017, volume=null, issue=null, pageStart=168, pageEnd=176, url=null, language=null, rfNumber=[8], rfOrder=7, authorNames=Shankaranarayana S M, Ram K, Mitra K, journalName=Fetal, Infant and Ophthalmic Medical Image Analysis:4th International Workshop, refType=null, unstructuredReference=Shankaranarayana S M, Ram K, Mitra K, et al. Joint optic disc and cup segmentation using fully convolutional and adversarial networks[C]// Fetal, Infant and Ophthalmic Medical Image Analysis:4th International Workshop. Berlin:Springer International Publishing, 2017: 168-176., articleTitle=Joint optic disc and cup segmentation using fully convolutional and adversarial networks, refAbstract=null), Reference(id=1204542863666290826, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[9], rfOrder=8, authorNames=Oktay O, Schemper J, Folgoc L L, journalName=arXiv preprint arXiv: 1804.03999, 2018, refType=null, unstructuredReference=Oktay O, Schemper J, Folgoc L L, et al. Attention Unet: learning where to look for the pancreas[J]. arXiv preprint arXiv: 1804.03999, 2018., articleTitle=Attention Unet: learning where to look for the pancreas, refAbstract=null), Reference(id=1204542863787925647, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, doi=null, pmid=null, pmcid=null, year=2018, volume=null, issue=null, pageStart=3, pageEnd=11, url=null, language=null, rfNumber=[10], rfOrder=9, authorNames=Zhou Z, Rahman S M M, Tajbakhsh N, journalName=Berlin:Springer International Publishing, refType=null, unstructuredReference=Zhou Z, Rahman S M M, Tajbakhsh N, et al. Unet++: A nested U-net architecture for medical image segmentation[C]//Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support:4th International Workshop. Berlin:Springer International Publishing, 2018: 3-11., articleTitle=Unet++: A nested U-net architecture for medical image segmentation, refAbstract=null), Reference(id=1204542863884394640, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, doi=null, pmid=null, pmcid=null, year=2017, volume=null, issue=null, pageStart=1, pageEnd=5, url=null, language=null, rfNumber=[11], rfOrder=10, authorNames=Gao X, Cai Y, Qiu C, journalName=2017 10th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics(CISP-BMEI), refType=null, unstructuredReference=Gao X, Cai Y, Qiu C, et al. Retinal blood vessel segmentation based on the Gaussian matched filter and U-net[C]// 2017 10th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics(CISP-BMEI). Ner York: IEEE, 2017: 1-5., articleTitle=Retinal blood vessel segmentation based on the Gaussian matched filter and U-net, refAbstract=null), Reference(id=1204542863985057941, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, doi=null, pmid=null, pmcid=null, year=2020, volume=392, issue=null, pageStart=268, pageEnd=276, url=null, language=null, rfNumber=[12], rfOrder=11, authorNames=Feng S, Zhuo Z, Pan D, journalName=Neurocomputing, refType=null, unstructuredReference=Feng S, Zhuo Z, Pan D, et al. CcNet: A cross-connected convolutional network for segmenting retinal vessels using multi-scale features[J]. Neurocomputing, 2020, 392: 268-276., articleTitle=CcNet: A cross-connected convolutional network for segmenting retinal vessels using multi-scale features, refAbstract=null), Reference(id=1204542864077332634, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, doi=null, pmid=null, pmcid=null, year=2019, volume=39, issue=5, pageStart=1392, pageEnd=1403, url=null, language=null, rfNumber=[13], rfOrder=12, authorNames=Mou L, Chen L, Cheng J, journalName=IEEE Transactions on Medical Imaging, refType=null, unstructuredReference=Mou L, Chen L, Cheng J, et al. Dense dilated network with probability regularized walk for vessel detection[J]. IEEE Transactions on Medical Imaging, 2019, 39(5): 1392-1403., articleTitle=Dense dilated network with probability regularized walk for vessel detection, refAbstract=null), Reference(id=1204542864161218718, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, doi=null, pmid=null, pmcid=null, year=2023, volume=23, issue=5, pageStart=1875, pageEnd=1881, url=null, language=null, rfNumber=[14], rfOrder=13, authorNames=黄伟安, 黎峰, 张雨, journalName=科学技术与工程, refType=null, unstructuredReference=黄伟安, 黎峰, 张雨, 等. 结合长短期记忆网络与Unet模型的心率估计方法[J]. 科学技术与工程, 2023, 23(5): 1875-1881., articleTitle=结合长短期记忆网络与Unet模型的心率估计方法, refAbstract=null), Reference(id=1204542864249299107, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, doi=null, pmid=null, pmcid=null, year=2023, volume=23, issue=5, pageStart=1875, pageEnd=1881, url=null, language=null, rfNumber=[14], rfOrder=14, authorNames=Huang Weian, Li Feng, Zhang Yu, journalName=Science Technology and Engineering, refType=null, unstructuredReference=Huang Weian, Li Feng, Zhang Yu, et al. Heart rate estimation method combining long short-term memory network and Unet model[J]. Science Technology and Engineering, 2023, 23(5): 1875-1881., articleTitle=Heart rate estimation method combining long short-term memory network and Unet model, refAbstract=null), Reference(id=1204542864320602280, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, doi=null, pmid=null, pmcid=null, year=2019, volume=null, issue=null, pageStart=1911, pageEnd=1920, url=null, language=null, rfNumber=[15], rfOrder=15, authorNames=Ding X, Guo Y, Ding G, journalName=Proceedings of the IEEE/CVF International Conference on Computer Vision, refType=null, unstructuredReference=Ding X, Guo Y, Ding G, et al. Acnet: Strengthening the kernel skeletons for powerful cnn via asymmetric convolution blocks[C]// Proceedings of the IEEE/CVF International Conference on Computer Vision. Ner York: IEEE, 2019: 1911-1920., articleTitle=Acnet: Strengthening the kernel skeletons for powerful cnn via asymmetric convolution blocks, refAbstract=null), Reference(id=1204542864421265584, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, doi=null, pmid=null, pmcid=null, year=1989, volume=null, issue=null, pageStart=239, pageEnd=null, url=null, language=null, rfNumber=[16], rfOrder=16, authorNames=Murenzi R, journalName=null, refType=null, unstructuredReference=Murenzi R. Wavelet transforms associated to the n-dimensional Euclidean group with dilations: signal in more than one dimension[M]// Wavelets:Time-Frequency Methods and Phase Space. Berlin: Springer Berlin Heidelberg, 1989: 239-246., articleTitle=Wavelet transforms associated to the n-dimensional Euclidean group with dilations: signal in more than one dimension, refAbstract=null), Reference(id=1204542864517734583, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, doi=null, pmid=null, pmcid=null, year=2021, volume=21, issue=22, pageStart=9461, pageEnd=9469, url=null, language=null, rfNumber=[17], rfOrder=17, authorNames=孙颢洋, 王欣, 曹昭睿, journalName=科学技术与工程, refType=null, unstructuredReference=孙颢洋, 王欣, 曹昭睿, 等. 基于深度学习的反无人机预警神经网络[J]. 科学技术与工程, 2021, 21(22): 9461-9469., articleTitle=基于深度学习的反无人机预警神经网络, refAbstract=null), Reference(id=1204542864714866879, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, doi=null, pmid=null, pmcid=null, year=2021, volume=21, issue=22, pageStart=9461, pageEnd=9469, url=null, language=null, rfNumber=[17], rfOrder=18, authorNames=Sun Haoyang, Wang Xin, Cao Zhaorui, journalName=Science Technology and Engineering, refType=null, unstructuredReference=Sun Haoyang, Wang Xin, Cao Zhaorui, et al. Neural network for early warning of anti-unmanned aerial vehicle based on deep learning[J]. Science Technology and Engineering, 2021, 21(22): 9461-9469., articleTitle=Neural network for early warning of anti-unmanned aerial vehicle based on deep learning, refAbstract=null), Reference(id=1204542864853278923, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, doi=null, pmid=null, pmcid=null, year=2018, volume=null, issue=null, pageStart=1451, pageEnd=1460, url=null, language=null, rfNumber=[18], rfOrder=19, authorNames=Wang P, Chen P, Yuan Y, journalName=null, refType=null, unstructuredReference=Wang P, Chen P, Yuan Y, et al.Understanding convolution for semantic segmentation[C]//2018 IEEE Winter Conference on Applications of Computer Vision(WACV). Ner York: IEEE, 2018: 1451-1460., articleTitle=null, refAbstract=null), Reference(id=1204542864953942227, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, doi=null, pmid=null, pmcid=null, year=2004, volume=23, issue=4, pageStart=501, pageEnd=509, url=null, language=null, rfNumber=[19], rfOrder=20, authorNames=Staal J, Abràmoff M D, Niemeijer M, journalName=IEEE Transactions on Medical Imaging, refType=null, unstructuredReference=Staal J, Abràmoff M D, Niemeijer M, et al. Ridge-based vessel segmentation in color images of the retina[J]. IEEE Transactions on Medical Imaging, 2004, 23(4): 501-509., articleTitle=Ridge-based vessel segmentation in color images of the retina, refAbstract=null), Reference(id=1204542865046216920, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, doi=null, pmid=null, pmcid=null, year=2015, volume=19, issue=1, pageStart=46, pageEnd=57, url=null, language=null, rfNumber=[20], rfOrder=21, authorNames=Azzopardi G, Strisciuglio N, Vento M, journalName=Medical Image Analysis, refType=null, unstructuredReference=Azzopardi G, Strisciuglio N, Vento M, et al. Trainable COSFIRE filters for vessel delineation with application to retinal images[J]. Medical Image Analysis, 2015, 19(1): 46-57., articleTitle=Trainable COSFIRE filters for vessel delineation with application to retinal images, refAbstract=null), Reference(id=1204542865155268828, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[21], rfOrder=22, authorNames=Chen Y, journalName=arXiv preprint arXiv: 1704.07502, 2017, refType=null, unstructuredReference=Chen Y. A labeling-free approach to supervising deep neural networks for retinal blood vessel segmentation[J]. arXiv preprint arXiv: 1704.07502, 2017., articleTitle=A labeling-free approach to supervising deep neural networks for retinal blood vessel segmentation, refAbstract=null), Reference(id=1204542865264320738, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, doi=null, pmid=null, pmcid=null, year=2019, volume=6, issue=1, pageStart=014006, pageEnd=null, url=null, language=null, rfNumber=[22], rfOrder=23, authorNames=Alom M Z, Yakopcic C, Hasan M, journalName=Journal of Medical Imaging, refType=null, unstructuredReference=Alom M Z, Yakopcic C, Hasan M, et al. Recurrent residual U-Net for medical image segmentation[J]. Journal of Medical Imaging, 2019, 6(1): 014006., articleTitle=Recurrent residual U-Net for medical image segmentation, refAbstract=null), Reference(id=1204542865356595428, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, doi=null, pmid=null, pmcid=null, year=2021, volume=null, issue=null, pageStart=1236, pageEnd=1242, url=null, language=null, rfNumber=[23], rfOrder=24, authorNames=Guo C, Szemenyei M, Yi Y, journalName=null, refType=null, unstructuredReference=Guo C, Szemenyei M, Yi Y, et al. Sa-unet:spatial attention U-net for retinal vessel segmentation[C]//2020 25th International Conference on Pattern Recognition(ICPR). New York: IEEE, 2021: 1236-1242., articleTitle=Sa-unet:spatial attention U-net for retinal vessel segmentation[C], refAbstract=null)], funds=[Fund(id=1204542861212622898, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, awardId=ZR2022MA027, language=CN, fundingSource=山东省自然科学基金面上项目(ZR2022MA027), fundOrder=null, country=null), Fund(id=1204542861338452023, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, awardId=2021003, language=CN, fundingSource=中山大学广东省计算科学重点实验室开放基金(2021003), fundOrder=null, country=null), Fund(id=1204542862462525497, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, awardId=2D-C-20190158, language=CN, fundingSource=山东新华医疗器械股份有限公司横向课题(2D-C-20190158), fundOrder=null, country=null), Fund(id=1204542862575771711, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, awardId=null, language=CN, fundingSource=山东省属普通本科高校教师访学研修项目, fundOrder=null, country=null)], companyList=[AuthorCompany(id=1204542855709696750, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, xref=null, ext=[AuthorCompanyExt(id=1204542855718085360, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, companyId=1204542855709696750, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1. School of Mathematics and Statistics, Shandong University of Technology, Zibo 255000, China), AuthorCompanyExt(id=1204542855726473969, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, companyId=1204542855709696750, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1.山东理工大学数学与统计学院, 淄博 255000)]), AuthorCompany(id=1204542855822942973, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, xref=null, ext=[AuthorCompanyExt(id=1204542855831331581, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, companyId=1204542855822942973, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2. Shinva Medical Instrument Co., Ltd., Zibo 255086, China), AuthorCompanyExt(id=1204542855839720190, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, companyId=1204542855822942973, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2.山东新华医疗器械股份有限公司, 淄博 255086)])], figs=[ArticleFig(id=1204542859291632587, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, language=EN, label=Fig.1, caption=CADCNet network structure, figureFileSmall=c+TaAvCUUzbNWDXUijPItg==, figureFileBig=oFmFfHfcXvZkVAZoscFepA==, tableContent=null), ArticleFig(id=1204542859396490193, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, language=CN, label=图1, caption=CADCNet网络结构, figureFileSmall=c+TaAvCUUzbNWDXUijPItg==, figureFileBig=oFmFfHfcXvZkVAZoscFepA==, tableContent=null), ArticleFig(id=1204542859543290843, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, language=EN, label=Fig.2, caption=AC-ConvNet module, figureFileSmall=bAtEP/omuNtoklGURWIVbg==, figureFileBig=G/UF+xcMrMVKym/Nm7oEdQ==, tableContent=null), ArticleFig(id=1204542859669119969, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, language=CN, label=图2, caption=AC-ConvNet模块, figureFileSmall=bAtEP/omuNtoklGURWIVbg==, figureFileBig=G/UF+xcMrMVKym/Nm7oEdQ==, tableContent=null), ArticleFig(id=1204542859761394664, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, language=EN, label=Fig.3, caption=Continuous dilation convolution module, figureFileSmall=SACBMSOjTxA4YJySEkCWRg==, figureFileBig=CnSC2QTME0IZ/NrBxHtF/g==, tableContent=null), ArticleFig(id=1204542859904001006, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, language=CN, label=图3, caption=连续膨胀卷积模块, figureFileSmall=SACBMSOjTxA4YJySEkCWRg==, figureFileBig=CnSC2QTME0IZ/NrBxHtF/g==, tableContent=null), ArticleFig(id=1204542860034024437, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, language=EN, label=Fig.4, caption=Continuous dilated convolution combination with different dilated rate, figureFileSmall=5YnjOWDBWv1TskDS1XyUiA==, figureFileBig=yB7biWAwgmS+TVvEP/qnbw==, tableContent=null), ArticleFig(id=1204542860168242172, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, language=CN, label=图4, caption=不同膨胀率的连续膨胀卷积组合, figureFileSmall=5YnjOWDBWv1TskDS1XyUiA==, figureFileBig=yB7biWAwgmS+TVvEP/qnbw==, tableContent=null), ArticleFig(id=1204542860285681664, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, language=EN, label=Fig.5, caption=Comparison of experimental results, figureFileSmall=QnzZIdSduq5loTTShR7Nfw==, figureFileBig=Gau58UryCYUrzT4TLZhrfg==, tableContent=null), ArticleFig(id=1204542860382150663, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, language=CN, label=图5, caption=实验结果对比, figureFileSmall=QnzZIdSduq5loTTShR7Nfw==, figureFileBig=Gau58UryCYUrzT4TLZhrfg==, tableContent=null), ArticleFig(id=1204542860482813963, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, language=EN, label=Table 1, caption=

Comparison of experimental result of different algorithms on the DRIVE dataset

, figureFileSmall=null, figureFileBig=null, tableContent=
方法 准确率 敏感度 特异性 F1
UNet[6] 0.949 2 0.720 8 0.972 7 0.815 2
UNet++[10] 0.950 4 0.728 6 0.975 6 0.818 9
Chen[21] 0.946 5 0.729 6 0.974 5 0.817 3
Res-UNet[8] 0.951 3 0.731 6 0.972 9 0.815 8
Alom等[22] 0.952 4 0.736 3 0.971 3 0.818 6
SA-UNet[23] 0.953 7 0.779 3 0.974 0 0.825 5
本文改进算法 0.956 8 0.764 3 0.973 2 0.832 6
), ArticleFig(id=1204542860545728528, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, language=CN, label=表1, caption=

DRIVE 数据集上各种算法的实验结果对比

, figureFileSmall=null, figureFileBig=null, tableContent=
方法 准确率 敏感度 特异性 F1
UNet[6] 0.949 2 0.720 8 0.972 7 0.815 2
UNet++[10] 0.950 4 0.728 6 0.975 6 0.818 9
Chen[21] 0.946 5 0.729 6 0.974 5 0.817 3
Res-UNet[8] 0.951 3 0.731 6 0.972 9 0.815 8
Alom等[22] 0.952 4 0.736 3 0.971 3 0.818 6
SA-UNet[23] 0.953 7 0.779 3 0.974 0 0.825 5
本文改进算法 0.956 8 0.764 3 0.973 2 0.832 6
), ArticleFig(id=1204542860667363352, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, language=EN, label=Table 2, caption=

Comparison of experimental result of different algorithms on the CHASE_DB1 dataset

, figureFileSmall=null, figureFileBig=null, tableContent=
方法 准确率 敏感度 特异性 F1
UNet[6] 0.948 9 0.722 3 0.972 4 0.814 3
UNet++[10] 0.951 2 0.729 6 0.974 6 0.817 6
Chen[21] 0.948 6 0.730 5 0.975 2 0.818 5
Res-UNet[8] 0.953 5 0.732 4 0.972 6 0.816 7
Alom等[22] 0.951 5 0.737 6 0.971 3 0.819 4
SA-UNet[23] 0.958 4 0.781 2 0.974 3 0.827 8
本文改进算法 0.959 8 0.766 2 0.973 1 0.830 4
), ArticleFig(id=1204542860763832347, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, language=CN, label=表2, caption=

CHASE_DB1 数据集上各种算法的实验结果对比

, figureFileSmall=null, figureFileBig=null, tableContent=
方法 准确率 敏感度 特异性 F1
UNet[6] 0.948 9 0.722 3 0.972 4 0.814 3
UNet++[10] 0.951 2 0.729 6 0.974 6 0.817 6
Chen[21] 0.948 6 0.730 5 0.975 2 0.818 5
Res-UNet[8] 0.953 5 0.732 4 0.972 6 0.816 7
Alom等[22] 0.951 5 0.737 6 0.971 3 0.819 4
SA-UNet[23] 0.958 4 0.781 2 0.974 3 0.827 8
本文改进算法 0.959 8 0.766 2 0.973 1 0.830 4
), ArticleFig(id=1204542860877078564, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, language=EN, label=Table 3, caption=

Comparison of evaluation result of the ablation experiments

, figureFileSmall=null, figureFileBig=null, tableContent=
方法 DRIVE数据集 CHASE_DB1数据集
准确率 F1 准确率 F1
UNet[6] 0.949 2 0.815 2 0.948 9 0.814 3
模型1 0.951 3 0.822 4 0.952 3 0.817 6
模型2 0.954 6 0.827 5 0.955 8 0.824 5
本文改进算法 0.956 8 0.832 6 0.959 8 0.830 4
), ArticleFig(id=1204542860952576042, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156907874640286636, language=CN, label=表3, caption=

消融实验客观评价结果

, figureFileSmall=null, figureFileBig=null, tableContent=
方法 DRIVE数据集 CHASE_DB1数据集
准确率 F1 准确率 F1
UNet[6] 0.949 2 0.815 2 0.948 9 0.814 3
模型1 0.951 3 0.822 4 0.952 3 0.817 6
模型2 0.954 6 0.827 5 0.955 8 0.824 5
本文改进算法 0.956 8 0.832 6 0.959 8 0.830 4
)], attaches=null, journal=Journal(id=1146119176004939786, delFlag=0, nameCn=科学技术与工程, nameEn=Science Technology and Engineering, nameHistory1=null, nameHistory2=null, issn=1671-1815, eissn=, cn=11-4688/T, coden=null, periodic=4, language=CN, oaType=是, ccby=null, superviseOffice=null, ownerOffice=null, pubOffice=null, editorOffice=null, officeType=null, aims=null, clcCode=null, officeProv=null, officeCity=null, officeAddr=null, officeZip=null, officeEmail=null, officePhone=null, editDirector=null, officeDirector=null, officeDirectorPhone=null, officeStaffNum=null, officeEmpNum=null, coverPicUrl=UKU/O7GSka5polgCTkbIIw==, journalPrice=null, startedYear=null, abbrevIsoEn=Sci Technol Eng, journalRemark=null, publicationField=null, createdTime=null, updatedTime=1754445529766, createdBy=null, updatedBy=13701087609, firstLetterCn=S, firstLetterEn=S, subjectCode=Natural Sciences, subjectName=自然科学, subjectCodeEn=Natural Sciences, subjectNameEn=null, picCn=UKU/O7GSka5polgCTkbIIw==, picEn=5hwlULoNwcbj3xUmVi9MAQ==, jcr=null, cjcr=null, exts=[JournalExt(id=1159791870395564357, language=CN, name=科学技术与工程, nameHistory1=null, nameHistory2=null, managedBy=, sponsoredBy=, publishedBy=, editorOffice=, officeProv=null, officeCity=null, officeAddr=, officeZip=, editDirector=null, officeDirector=null, officePhone=null, coverPicUrl=null, journalRemark=, submitArticleUrl=null, websiteUrl=http://www.stae.com.cn/jsygc/home, createdTime=1754445529793, updatedTime=1754445529793, createdBy=13701087609, updatedBy=13701087609, submissionGuidelinesUrl=http://www.stae.com.cn/jsygc/site/menus/20090429150146001, submissionAuthorUrl=http://www.stae.com.cn/jsygc/author/login, submissionEditorUrl=http://www.stae.com.cn/jsygc/editor/login, submissionReviewUrl=http://www.stae.com.cn/jsygc/reviewer/login, submissionCeEditorUrl=, submissionAeEditorUrl=, option={"copyright":""}), JournalExt(id=1159791870441701702, language=EN, name=Science Technology and Engineering, nameHistory1=null, nameHistory2=null, managedBy=, sponsoredBy=, publishedBy=, editorOffice=, officeProv=null, officeCity=null, officeAddr=, officeZip=, editDirector=null, officeDirector=null, officePhone=null, coverPicUrl=null, journalRemark=, submitArticleUrl=null, websiteUrl=http://www.stae.com.cn/jsygc/home, createdTime=1754445529804, updatedTime=1754445529804, createdBy=13701087609, updatedBy=13701087609, submissionGuidelinesUrl=, submissionAuthorUrl=http://www.stae.com.cn/jsygc/author/login, submissionEditorUrl=http://www.stae.com.cn/jsygc/editor/login, submissionReviewUrl=http://www.stae.com.cn/jsygc/reviewer/login, submissionCeEditorUrl=, submissionAeEditorUrl=, option={"copyright":""})], databaseList=null, tenantJournalId=1146123166801305609, websiteList=[Website(id=1148243202391400884, webName=null, webTitle=null, webDomain=null, webCopyrigh=null, webIpcNo=null, seoTitle=null, seoKeywords=null, seoDescription=null, tenantJournalId=null, journalId=1146123166801305609, journalNameCn=null, journalNameEn=null, grayFlag=null, tenantId=1146029695717560320, platformId=null, journalGroupId=null, journalGroupNameCn=null, journalGroupNameEn=null, type=1, domain=https://castjournals.cast.org.cn/joweb/kxjsygc/CN, language=CN, createTime=1751692112777, createBy=18614031015, updateTime=1753520965431, updateBy=18614031015, name=科学技术与工程-中文站点, tplId=1146099689490845704, title=科学技术与工程, delFlag=0, indexPage=/home, props=[WebsiteProps(id=1148622798802673703, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1148243202391400884, code=articleTextType, value=kx, createTime=1751782615614, updateTime=1751782615614, creator=18614031015, updator=18614031015), WebsiteProps(id=1148622798781702180, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1148243202391400884, code=banner, value=null, createTime=1751782615609, updateTime=1751782615609, creator=18614031015, updator=18614031015), WebsiteProps(id=1148622798769119267, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1148243202391400884, code=logo, value=https://castjournals.cast.org.cn/joweb/kjdb/CN/file/pic?fileId=j86gbwi+p0Idkyl5SzIlmQ==, createTime=1751782615606, updateTime=1751782615606, creator=18614031015, updator=18614031015), WebsiteProps(id=1148622798794285094, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1148243202391400884, code=picServerUrl, value=https://castjournals.cast.org.cn/joweb/kjdb/CN/file/pic, createTime=1751782615612, updateTime=1751782615612, creator=18614031015, updator=18614031015), WebsiteProps(id=1148622798790090789, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1148243202391400884, code=staticResourcePath, value=https://castjournals.cast.org.cn/joweb/cast_kjdb_cn_619/, createTime=1751782615611, updateTime=1751782615611, creator=18614031015, updator=18614031015)]), Website(id=1155914124811976731, webName=null, webTitle=null, webDomain=null, webCopyrigh=null, webIpcNo=null, seoTitle=null, seoKeywords=null, seoDescription=null, tenantJournalId=null, journalId=1146123166801305609, journalNameCn=null, journalNameEn=null, grayFlag=null, tenantId=1146029695717560320, platformId=null, journalGroupId=null, journalGroupNameCn=null, journalGroupNameEn=null, type=1, domain=https://castjournals.cast.org.cn/joweb/kxjsygc/EN, language=EN, createTime=1753521003206, createBy=18614031015, updateTime=1753521003206, updateBy=18614031015, name=科学技术与工程-英文站点, tplId=1146101810881728533, title=Science Technology and Engineering, delFlag=0, indexPage=/home, props=[WebsiteProps(id=1155914371227308235, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1155914124811976731, code=articleTextType, value=kx, createTime=1753521061952, updateTime=1753521061952, creator=18614031015, updator=18614031015), WebsiteProps(id=1155914371210531016, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1155914124811976731, code=banner, value=null, createTime=1753521061947, updateTime=1753521061947, creator=18614031015, updator=18614031015), WebsiteProps(id=1155914371202142407, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1155914124811976731, code=logo, value=https://castjournals.cast.org.cn/joweb/kjdb/CN/file/pic?fileId=j86gbwi+p0Idkyl5SzIlmQ==, createTime=1753521061945, updateTime=1753521061945, creator=18614031015, updator=18614031015), WebsiteProps(id=1155914371223113930, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1155914124811976731, code=picServerUrl, value=https://castjournals.cast.org.cn/joweb/kjdb/CN/file/pic, createTime=1753521061950, updateTime=1753521061950, creator=18614031015, updator=18614031015), WebsiteProps(id=1155914371218919625, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1155914124811976731, code=staticResourcePath, value=https://castjournals.cast.org.cn/joweb/cast_kjdb_cn_619/, createTime=1753521061949, updateTime=1753521061949, creator=18614031015, updator=18614031015)])], journalTitle=科学技术与工程, weixinUrl=null, journalUrl=null, iacademicId=null, status=0, seqNo=null, journalTitleEn=Science Technology and Engineering, journalPhotoCn=UKU/O7GSka5polgCTkbIIw==, journalPhotoEn=5hwlULoNwcbj3xUmVi9MAQ==, journalFirstLetter=S, journalRecommend=null, journalNew=null, journalCollection=null, jcrJf=null, cjcrJf=null, jcrJfStr=null, cjcrJfStr=null, submissionFirstDecision=null, sciSubjectClassification=null, casSubjectClassification=null, citeScore=null, totalCitationFrequency=null, icpCode=null, psCode=null, advertisingLicenseCode=null, copyrightInformation=null, country=null, option=null, provinceCode=null, provinceName=null, collectFlag=false), detailUrlCn=https://castjournals.cast.org.cn/joweb/kxjsygc/CN/10.12404/j.issn.1671-1815.2309339, detailUrlEn=https://castjournals.cast.org.cn/joweb/kxjsygc/EN/10.12404/j.issn.1671-1815.2309339, pdfUrlCn=https://castjournals.cast.org.cn/joweb/kxjsygc/CN/PDF/10.12404/j.issn.1671-1815.2309339, pdfUrlEn=https://castjournals.cast.org.cn/joweb/kxjsygc/EN/PDF/10.12404/j.issn.1671-1815.2309339, aliStartDate=null, aliEndDate=null, collectionFlag=false, citedCount=null, citedUrl=null, reference=null)
收藏切换
CADCNet: 一种改进的视网膜血管分割算法
收藏切换
PDF下载
岳昱超 1 , 王迎美 1, * , 秦嘉川 2
科学技术与工程 | 论文·医药、卫生 2025,25(3): 962-968
收起
收藏切换
科学技术与工程 | 论文·医药、卫生 2025, 25(3): 962-968
CADCNet: 一种改进的视网膜血管分割算法
全屏
岳昱超1 , 王迎美1, * , 秦嘉川2
作者信息
  • 1.山东理工大学数学与统计学院, 淄博 255000
  • 2.山东新华医疗器械股份有限公司, 淄博 255086
  • 岳昱超(1998—),男,汉族,山东威海人,硕士研究生。研究方向:医学图像处理。E-mail:

通讯作者:

* 王迎美(1987—),女,汉族,山东新泰人,博士,副教授。研究方向:医学图像处理与重建。E-mail:
CADCNet: An Improved Algorithm for Retinal Vessel Segmentation
Yu-chao YUE1 , Ying-mei WANG1, * , Jia-chuan QIN2
Affiliations
  • 1. School of Mathematics and Statistics, Shandong University of Technology, Zibo 255000, China
  • 2. Shinva Medical Instrument Co., Ltd., Zibo 255086, China
出版时间: 2025-01-28 doi: 10.12404/j.issn.1671-1815.2309339
文章导航
收藏切换

目前传统的视网膜血管分割方法存在的视盘混淆引起的误分割、分割结果缺乏连续性,以及细节区域分割不精准等问题。为解决这一难题,提出了一种基于UNet的视网膜血管分割算法。该算法利用两个水平和垂直一维卷积和二维方形卷积的融合替代传统方形卷积,提高了眼球区域的表征能力;采用了多尺度分支增加特征空间的多样性,提升了网络的特征学习和表达能力。此外,为进一步改善分割效果,还将多层膨胀卷积引入自编码器的深层结构,替代了传统的简单池化操作,增大卷积核的大小,扩大了感受野范围,实现了多尺度浅层特征和深层特征信息的融合。本文算法在公开DRIVE和CHASE_DB1两个数据集上进行了评估,实验结果表明,本文算法的精确率和F1上分别达到了0.956 8、0.959 8和0.832 6、0.830 4。与传统的UNet和近期部分UNet改进网络视网膜血管分割方法相比,本文算法在准确率、敏感度、特异性、F1指标上表现出一定的优势,这一验证结果充分证明了本文所提出的模型在分割任务上具备较强的精确分割能力。

视网膜血管分割  /  连续膨胀卷积  /  深度学习  /  不对称卷积  /  UNet模型

Traditional retinal vessel segmentation methods often face challenges such as missegmentation caused by optic disc confusion, lack of continuity in segmentation results, and imprecise segmentation in detailed regions. To address these issues, a retinal vessel segmentation algorithm was proposed based on UNet. The algorithm replaced traditional square convolutions with a fusion of horizontal and vertical one-dimensional convolutions and two-dimensional square convolutions, enhancing the representation capability of the eye region. A multi-scale branch approach was adopted to increase feature space diversity, thereby improving the network’s feature learning and expression capabilities. Additionally, to further enhance segmentation performance, multi-layer dilated convolutions was introduced into the deep structure of the autoencoder, replacing traditional simple pooling operations. This approach enlarged the convolution kernel size and expanded the receptive field, achieving a fusion of multi-scale shallow and deep feature information. The proposed algorithm was evaluated on the public DRIVE and CHASE_DB1 datasets. Experimental results demonstrates that the algorithm achieves precision (0.956 8 and 0.959 8) and F1 scores (0.832 6 and 0.830 4), respectively. Compared with traditional UNet and recent UNet-based retinal vessel segmentation methods, the proposed algorithm shows advantages in accuracy, sensitivity, specificity, and F1 metrics, these validation results fully demonstrate the proposed model’s strong capability in precise segmentation tasks.

retinal vessel segmentation  /  continuous dilation convolution  /  deep learning  /  asymmetric convolution  /  UNet model
岳昱超, 王迎美, 秦嘉川. CADCNet: 一种改进的视网膜血管分割算法. 科学技术与工程, 2025 , 25 (3) : 962 -968 . DOI: 10.12404/j.issn.1671-1815.2309339
Yu-chao YUE, Ying-mei WANG, Jia-chuan QIN. CADCNet: An Improved Algorithm for Retinal Vessel Segmentation[J]. Science Technology and Engineering, 2025 , 25 (3) : 962 -968 . DOI: 10.12404/j.issn.1671-1815.2309339
眼底图像处理技术在帮助医生诊断眼部疾病方面发挥着重要作用,包括老年黄斑病变[1]、病理性近视[2]、青光眼[3]和白内障[4]等。及早发现眼部病变可以使患者及时接受治疗,从而避免永久性视力损害。目前,医学专家主要通过观察视网膜血管的分布来进行疾病分析和诊断。然而,这种方法效率较低,而且人工勾画和观察血管分布存在主观性较强、工作量大的问题。因此,利用计算机辅助诊断进行视网膜血管的自动检测与分割具有重要的研究和应用价值。
在计算机视觉领域,深度学习技术展现出了巨大的优势,人工神经网络已被广泛应用于图像分类、分割和识别等任务。针对医学图像分割方法,许多研究者进行了相关研究。Long等[5]首次提出全卷积网络(fully convolutional network,FCN)网络。自从其问世以来,FCN已成为像素级预测的首个端到端网络架构,在语义分割领域扮演着基础框架的角色。随后的算法主要是在这一网络结构基础上的改进。Ronneberger等[6]提出了U形神经网络UNet用于生物医学图像分割,该网络通过跳跃连接在解码器中融合高低层的特征信息,增强了局部细节恢复能力,效果显著。由于其出色的分割性能,UNet逐渐成为医学分割领域的热点,随后出现了许多基于UNet的改进网络模型。
为了提高图像分割的准确性,Yu等[7]提出了一种基于残差网络的创新网络结构,采用了膨胀卷积(也称为“空洞卷积”)以拓展感受野,从而显著增强了网络对于大尺度特征的学习能力。Shankaranarayana等[8]结合了残差模块和UNet的思想,提出了Res-UNet。残差模块作为深度残差网络中的基本构建单元,用于解决深层网络训练过程中的梯度消失和梯度爆炸问题。这一改进在UNet的基础上进一步提升了图像分割的性能。Oktay等[9]提出了Attention UNet,这一方法在传统的UNet架构中引入了注意力机制,将注意力机制应用于UNet分割网络中,可以更好地捕捉显著特征并抑制无关背景区域。另外,Zhou等[10]提出了UNet++作为对UNet的改进版本。UNet++通过嵌套和稠密的跳跃连接来增强特征表示和分割准确性,并设计了一种剪枝策略以提高推理速度,同时保持性能。
虽然目前存在许多图像分割模型,然而在不同数据集上它们的分割效果存在很大差异。在医学影像分割领域,UNet网络因其独特而简洁的对称设计已成为许多学者进行研究的基础架构。在视网膜血管分割领域中,Gao等[11]将视网膜血管分割任务视为多标签问题,结合UNet网络与高斯滤波在对比度不足的区域和病理性区域能够精确地分离血管与背景。Feng等[12]提出了一种自动分割视网膜血管的交叉连接卷积神经网络(cross-connected convolutional neural network,CcNet),通过主路径和次路径之间的交叉连接融合多级特征,实现了对多尺度信息的自动有效学习。Mou等[13]专注于血管的连通性,采用稠密扩张网络进行初始检测,使用概率正规化行走算法解决了初始检测中的断裂问题。此外,在医学分割中的其他领域,如心率估计[14]等领域,UNet网络也展现出令人满意的效果。
结合UNet网络的优点和专家改进UNet网络以适应不同分割任务的思路,针对视网膜血管结构复杂、细支血管分割困难等问题,提出一种基于UNet改进的CADCNet(continuous asymmetric dilation convolution network)分割网络。该方法在原始UNet的基础上提出了以下改进措施以提高网络的分割能力和对图像细节区域的分割效果:在编码器上引入ACNet (asymmetric convolution net work) 模块代替传统的双卷积模块,通过利用垂直、水平和方形卷积提取不同尺度的特征并将其融合,以获取更丰富的特征表示;在下采样过程中引入连续膨胀卷积模块,扩大了感受野,进一步增强了对细小血管的分割能力。实验结果表明,相比传统的UNet和近期部分UNet改进网络,CADCNet模型在处理复杂血管分割方面展现出显著的优势。
近年来,对于卷积神经网络(convolutional neural networks, CNN)的改进主要集中在以下几个方面:①改变CNN层之间的连接方式,如直接简单堆叠,类似于Res-Net的残差连接或DenseNet的密集连接;②对不同层的输出融合,例如,Inception家族的方法;③构建小模块与CNN结合,例如,在SENet(squeeze-and-excitation networks)中引入的SE模块,直接与当前的CNN结合。
通过这些改进,CNN可以以多种方式进行组合,构建出特定的CNN网络,从而提高模型性能。然而,这种通过增加可训练参数和复杂的连接来提升模型性能会增大计算量,且效果也不一定是越复杂越好。构建网络时,在不过分增加模型大小的前提下提高网络性能,也是一种较好的思路。Ding等[15]提出了ACNet,该网络在卷积核的设计上,不同于传统方形,将方形卷积核分解为竖条卷积核和横条卷积核,并将其与方形卷积核相结合。这种做法与两个方形卷积相比不会大量增加计算负担,同时可以增强标准卷积层的性能。
膨胀卷积最初是在小波分解算法[16]中引入的。膨胀卷积的核心思想是在卷积核的像素之间插入“洞”(即隔行隔列进行零填充),以实现在深度卷积神经网络中的密集特征提取,并提高图像分辨率。此外,膨胀卷积还能扩大感受野范围。近年来,膨胀卷积已经被广泛应用于各种任务,如故障诊断[17]、语义分割[18]等。
基于UNet框架构建了一种新的卷积神经网络结构,该结构由编码器和解码器组成,如图1所示。编码器包括图像输入、四次AC-ConvNet (asymmetric convolution convnet)卷积和四次下采样过程,不同于原始UNet的双方形卷积,每个卷积步骤中都会经历一次AC-ConvNet卷积模块。为了提高模型的鲁棒性和避免过拟合,AC-ConvNet模块中加入批标准化(batch normalization,BN)和ReLu激活函数。下采样过程中的前三次下采样使用最大池化来获得深层次语义信息。不同于前三次下采样,在该模型的第四次下采样阶段,引入连续的膨胀卷积层(continuous dilation convolution,CDC)替代传统的最大池化层,通过扩大像素感受野区域的方式解决了血管末端分割中存在的不连续性问题,保留了更多的细节信息,有效提升血管分割的准确性和连续性。解码器包括四次AC-ConvNet卷积、四次上采样、全连接层和图像输出过程,解码阶段每一步骤都经过一个AC-ConvNet模块、上采样和对应编码阶段图像连接,最后通过全连接层及Softmax激活函数输出视网膜血管图像。
基于ACNet的思想,提出AC-ConvNet模块,替代传统的双重方形卷积,以强化主干网络,通过采用这种新的网络结构,能够提取更多视网膜血管细节区域的信息,从而提高血管分割的质量。该网络结构的设计旨在帮助模型更好地挖掘和保留血管细节信息与空间尺度信息,从而实现精确的分割结果。AC-ConvNet模块结构如图2所示,AC-ConvNet模块首先组合3个并行的3×1卷积、3×3卷积和1×3卷积模块,每个卷积后都加入BN和ReLu激活函数,使得AC-ConvNet具有非线性映射学习能力并且可以避免过拟合;然后,将各个并行的结果合并后得到中间层,然后接上一个3×3常规方形卷积,加入BN和ReLu激活函数最后进行输出。通过融合3个不同形状卷积核的特征图,能够获取更全面、更丰富的特征信息。这种结果的合并能够有效地提升语义分割任务的准确性和细节保留能力。
在网络的实现过程中,将编码阶段的每层双重卷积模块用AC-ConvNet模块代替。对比传统的双重卷积,AC-ConvNet模块在训练时增强了特征提取能力,提升了分割结果的质量。
对于细小血管图像而言,使用最大池化进行下采样会导致严重的细节丢失。为了提高分割精度的同时保持计算成本,在UNet的深层结构中使用膨胀卷积替代最大池化。然而,使用膨胀卷积时,随着膨胀率的增加,卷积核中的有效权重会减少。也就是说,在两个采样像素之间插入的零值越多,卷积的稀疏性就越高。当膨胀率足够大时,卷积会变得过于稀疏,从而失去信息获取的能力,并无法获取到全局的有用信息,这会影响模型的分割能力。过度稀疏的采样不利于卷积学习,导致局部信息完全丢失或采集的信息之间相隔过远,最终损害了信息的一致性。
为了应对该问题,提出CDC模块,如图3所示,在不增加计算开销的情况下解决该问题。该模块通过选择不同的膨胀率,将连续的3个膨胀卷积核融合在一起,以实现感受野的全覆盖。CDC模块的引入有效地避免了采样稀疏的影响,使得网络能够充分聚合多尺度的上下文信息,进而显著提升对视网膜细小血管的分割效果。具体而言,CDC模块的结构如图3所示。
在感受野大小确定的情况下,CDC模块中膨胀卷积的膨胀率选择具有重要意义。为了确保最终感受野能够完全覆盖目标区域且不出现空洞或边缘丢失的情况,连续的膨胀卷积的膨胀率需要经过合理的选择。下面通过目标感受野为13展示两个连续膨胀卷积组合的实际感受野情况,如图4所示。图4中的三列分别代表进行三次连续膨胀卷积后的感受野范围。第一行组合膨胀率为[2,2,2],第二行组合膨胀率为[1,2,3]。通过观察实验结果,可以发现在具有相同目标感受野的情况下,第一张图像未能实现感受野的全覆盖,而第二张图像成功实现了感受野的全覆盖。基于此,为了实现目标感受野全覆盖,要选择组合膨胀率为[1,2,3]。这样在目标感受野大小确定的情况下,能够充分覆盖整个目标区域。
采用DRIVE数据集[19]和CHASE_DB1数据集[20],DRIVE数据集是为深入研究视网膜病变而专门构建的数据库。其主要目标在于完成眼底图像中血管分割的任务。利用这一数据集,研究人员可以评估和比较眼底图像血管分割算法的效果。DRIVE数据集最初由荷兰图像科学研究所的Niemeijer团队于2004年发布。这一数据集由医院眼科专家进行手动标注,共包含40张像素为565×584的彩色眼底图像,训练集和测试集两个各20张。每张图像均配备了相应的二进制掩膜,便于准确标记血管区域。CHASE_DB1数据集用于眼底图像分析,这个数据集包含了从28名受试者的左眼和右眼中获取的28幅眼底图像,共56幅图像。每幅图像的分辨率为999×960像素。该数据集中的每幅眼底图像都配有对应的血管分割标注,这些标注是由专业的医学人员手工标记的。
为了增强血管结构的可见性并解决样本数量不足导致的过拟合问题,在训练前对数据进行一系列预处理操作,对数据集采用了镜像翻转、几何平移、随机缩放等数据增强方法进行了数据扩充。最终,通过裁剪操作得到了160张尺寸为480×480的图像作为训练集。通过数据集扩充,能够增加样本的多样性和数量,同时提升模型的泛化能力和鲁棒性。
本文模型是在Windows环境下实现的,并使用了以下软件环境:Pycharm平台、基于PyTorch框架的实现、Python3.9版本。实验所用电脑硬件环境为Windows 10系统,搭载Intel(R) Core(TM) i7-10875H CPU @2.30 GHz处理器和NVIDIA GeForce RTX 2060显卡。为了进行参数优化,采用了Adam优化器,并将学习率设置为1×10-4,权重衰减系数设置为0.1,以及批大小设置为4。
在本文中,采用交叉熵作为损失函数。交叉熵损失函数是一种常用的衡量两个概率分布之间差异的度量方式,用于评估模型预测结果与真实标签之间的不一致程度。交叉熵损失函数的数学公式为
yi=- i = 1 N [ y i l o g 2 p i + ( 1 - y i ) l o g 2 ( 1 - p i ) ] N
式(1)中:N为总样本数;Pi表示第i个样本预测为正的概率; yi表示第i个样本的标签。
为了验证本文模型算法的有效性,采用图像分割领域常用的评估标准,其中包括准确率A(accuracy)、 敏感度Sn (sensitivity)、特异度Sp(specificity)和 F1 值。
准确率是用于评估分类模型性能的指标之一,它表示分类模型正确预测样本的比例。准确率的计算公式为
A= T P + T N T P + T N + F P + F N
式(2)中:TP (true positive)表示正确分割的血管像素;FP (false positive)表示背景像素误分割为血管像素;TN (true negative)正确分割的背景像素;FN (false negative)表示血管像素误分成背景像素。
敏感度,也被称为真阳性率(true positive rate, TPR)或召回率R,用于衡量分类模型对正样本的识别能力。敏感度表示分类模型正确预测为正样本的比例。敏感度的计算公式为
Sn= T P T P + F N
特异度用于衡量分类模型对负样本的识别能力。特异度表示分类模型正确预测为负样本的比例。特异度的计算公式为
Sp= T N T N + F P
F1是综合考虑了准确率和敏感度的指标,用于评估分类模型的综合性能,是准确率和敏感度的调和平均值,计算公式为
F1= 2 T P 2 T P + F N + F P
F1是一种常用的评估模型性能的指标,其取值范围在0~1。F1较接近1表示模型的性能更佳。F1综合考虑了模型的准确率和敏感度,使其适用于处理不平衡数据集或对正负样本重视程度不同的情况。
使用提出的算法与UNet进行了分割实验,并给出了图像和分割指标的对比实验结果。如图5所示,第一列到第四列分别展示了原始眼底图像、UNet血管分割结果、改进算法血管分割结果和专家手工标记的标准分割结果图。前三行为测试集选取图片,最后一行为第三行红色方框放大图。从图5中的分割结果可以观察到,本文提出的算法与专家手工标注结果基本一致,相较于UNet的分割效果,在眼底血管分割中克服了难点,即使在局部细小血管边缘区域,仍能清晰地表现出血管的分割轮廓。此外,在主血管分割过程中,几乎没有断点,能够更好地捕捉细节信息,从而在视网膜血管分割任务中能取得更好的效果。
表1表2所示为本文提出的算法与近几年多种改进的UNet视网膜分割算法在公开的DRIVE数据集上的分割指标。从分割指标上可以看出,本文提出的方法在准确率、敏感度、特异性和F1指标上相较于近年来的对比算法基本都有所提高。在DRIVE和CHASE_DB1数据集上,本文方法的准确率达到了0.956 8和0.959 8,F1达到了0.832 6和0.830 4,均高于其他算法。
表3中,为了验证每个模块对本文提出算法的提升效果,对DRIVE数据集和CHASE_DB1数据集进行了消融实验。为了方便起见,把本文提出的模型去掉CDC模块后命名为模型1;去掉AC-ConvNet模块后命名为模型2,实验结果表明分割效果均有所下降,这充分证明了本文提出算法的有效性。
针对视网膜血管结构复杂、图像对比度低、细节区域分割不精问题,基于UNet架构提出了一种CADCNet网络模型,用AC-ConvNet卷积层替换了传统的双重卷积层,并提出了CDC模块进一步提升细微血管分割能力。实验结果表明,这些改进在深度特征提取和细节保持上的有效性。通过对数据集进行对比实验,观察到本文所提出的CADCNet模型相较于UNet模型和部分改进UNet模型在视网膜血管分割任务中在分割血管的连续性和细微血管的精确分割方面表现优越。AC-ConvNet卷积层的加入使网络能够捕捉更多细节信息,提升了对视网膜血管的准确分割能力。此外,最后一层的连续膨胀卷积拥有更大的感受野,能够更好地捕捉视网膜血管的一致性信息。
在视觉上,本文算法得到的血管分割结果呈现出清晰而明确的血管边界,几乎没有断裂和噪声伪影。该算法成功地解决了细小血管分割的问题,能够有效捕捉局部血管边缘的复杂性。总的来说,实验结果证实所提出的改进UNet方法在视网膜血管分割中的有效性。它成功地学习和融合了细节信息,提升了分割性能。综上所述,本文算法在视网膜血管分割中取得了良好的效果,具有一定的临床应用价值。
  • 山东省自然科学基金面上项目(ZR2022MA027)
  • 中山大学广东省计算科学重点实验室开放基金(2021003)
  • 山东新华医疗器械股份有限公司横向课题(2D-C-20190158)
  • 山东省属普通本科高校教师访学研修项目
参考文献 引证文献
排序方式:
[1]
Wong W L, Su X, Li X, et al. Global prevalence of age-related macular degeneration and disease burden projection for 2020 and 2040: a systematic review and meta-analysis[J]. The Lancet Global Health, 2014, 2(2): e106-e116.
[2]
Holden B A, Fricke T R, Wilson D A T R, et al. Global prevalence of myopia and high myopia and temporal trends from 2000 through 2050[J]. Ophthalmology, 2016, 123(5): 1036-1042.
[3]
Tham Y C, Li X, Wong T Y, et al. Global prevalence of glaucoma and projections of glaucoma burden through 2040: a systematic review and meta-analysis[J]. Ophthalmology, 2014, 121(11): 2081-2090.
[4]
Pascolini D, Mariotti S P. Global estimates of visual impairment: 2010[J]. British Journal of Ophthalmology, 2012, 96(5): 614-618.
[5]
Long J, Shelhamer E, Darrell T. Fully convolutional networks for semantic segmentation[C]// Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. New York: IEEE, 2015: 3431-3440.
[6]
Ronneberger O, Fischer P, Brox T. U-net: Convolutional networks for biomedical image segmentation[C]// Medical Image Computing and Computer-Assisted Intervention-MICCAI 2015: 18th International Conference. Berlin:Springer International Publishing, 2015: 234-241.
[7]
Yu F, Koltun V, Funkhouser T. Dilated residual networks[C]// Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. New York: IEEE, 2017: 472-480.
[8]
Shankaranarayana S M, Ram K, Mitra K, et al. Joint optic disc and cup segmentation using fully convolutional and adversarial networks[C]// Fetal, Infant and Ophthalmic Medical Image Analysis:4th International Workshop. Berlin:Springer International Publishing, 2017: 168-176.
[9]
Oktay O, Schemper J, Folgoc L L, et al. Attention Unet: learning where to look for the pancreas[J]. arXiv preprint arXiv: 1804.03999, 2018.
[10]
Zhou Z, Rahman S M M, Tajbakhsh N, et al. Unet++: A nested U-net architecture for medical image segmentation[C]//Deep Learning in Medical Image Analysis and Multimodal Learning for Clinical Decision Support:4th International Workshop. Berlin:Springer International Publishing, 2018: 3-11.
[11]
Gao X, Cai Y, Qiu C, et al. Retinal blood vessel segmentation based on the Gaussian matched filter and U-net[C]// 2017 10th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics(CISP-BMEI). Ner York: IEEE, 2017: 1-5.
[12]
Feng S, Zhuo Z, Pan D, et al. CcNet: A cross-connected convolutional network for segmenting retinal vessels using multi-scale features[J]. Neurocomputing, 2020, 392: 268-276.
[13]
Mou L, Chen L, Cheng J, et al. Dense dilated network with probability regularized walk for vessel detection[J]. IEEE Transactions on Medical Imaging, 2019, 39(5): 1392-1403.
[14]
黄伟安, 黎峰, 张雨, 等. 结合长短期记忆网络与Unet模型的心率估计方法[J]. 科学技术与工程, 2023, 23(5): 1875-1881.
Huang Weian, Li Feng, Zhang Yu, et al. Heart rate estimation method combining long short-term memory network and Unet model[J]. Science Technology and Engineering, 2023, 23(5): 1875-1881.
[15]
Ding X, Guo Y, Ding G, et al. Acnet: Strengthening the kernel skeletons for powerful cnn via asymmetric convolution blocks[C]// Proceedings of the IEEE/CVF International Conference on Computer Vision. Ner York: IEEE, 2019: 1911-1920.
[16]
Murenzi R. Wavelet transforms associated to the n-dimensional Euclidean group with dilations: signal in more than one dimension[M]// Wavelets:Time-Frequency Methods and Phase Space. Berlin: Springer Berlin Heidelberg, 1989: 239-246.
[17]
孙颢洋, 王欣, 曹昭睿, 等. 基于深度学习的反无人机预警神经网络[J]. 科学技术与工程, 2021, 21(22): 9461-9469.
Sun Haoyang, Wang Xin, Cao Zhaorui, et al. Neural network for early warning of anti-unmanned aerial vehicle based on deep learning[J]. Science Technology and Engineering, 2021, 21(22): 9461-9469.
[18]
Wang P, Chen P, Yuan Y, et al.Understanding convolution for semantic segmentation[C]//2018 IEEE Winter Conference on Applications of Computer Vision(WACV). Ner York: IEEE, 2018: 1451-1460.
[19]
Staal J, Abràmoff M D, Niemeijer M, et al. Ridge-based vessel segmentation in color images of the retina[J]. IEEE Transactions on Medical Imaging, 2004, 23(4): 501-509.
[20]
Azzopardi G, Strisciuglio N, Vento M, et al. Trainable COSFIRE filters for vessel delineation with application to retinal images[J]. Medical Image Analysis, 2015, 19(1): 46-57.
[21]
Chen Y. A labeling-free approach to supervising deep neural networks for retinal blood vessel segmentation[J]. arXiv preprint arXiv: 1704.07502, 2017.
[22]
Alom M Z, Yakopcic C, Hasan M, et al. Recurrent residual U-Net for medical image segmentation[J]. Journal of Medical Imaging, 2019, 6(1): 014006.
[23]
Guo C, Szemenyei M, Yi Y, et al. Sa-unet:spatial attention U-net for retinal vessel segmentation[C]//2020 25th International Conference on Pattern Recognition(ICPR). New York: IEEE, 2021: 1236-1242.
2025年第25卷第3期
PDF下载
345
147
引用本文
BibTeX
文章信息
doi: 10.12404/j.issn.1671-1815.2309339
  • 接收时间:2023-11-27
  • 首发时间:2025-07-29
  • 出版时间:2025-01-28
补充材料
相关文章
文章信息
作者
出版历史
  • 收稿日期:2023-11-27
  • 修回日期:2024-06-24
基金
山东省自然科学基金面上项目(ZR2022MA027)
中山大学广东省计算科学重点实验室开放基金(2021003)
山东新华医疗器械股份有限公司横向课题(2D-C-20190158)
山东省属普通本科高校教师访学研修项目
作者信息
    1.山东理工大学数学与统计学院, 淄博 255000
    2.山东新华医疗器械股份有限公司, 淄博 255086

通讯作者:

* 王迎美(1987—),女,汉族,山东新泰人,博士,副教授。研究方向:医学图像处理与重建。E-mail:
参考文献
分享链接
https://castjournals.cast.org.cn/joweb/kxjsygc/CN/10.12404/j.issn.1671-1815.2309339
分享至
全文二维码

扫描看全文

引用本文
BibTeX
本文的引用情况
2种不同金属材料的力学参数

Family
属数
Number of
genus
种数
Number of
species
占总种数比例
Percentage of
total species (%)

Genus
种数
Number of
species
占总种数比例
Percentage of total
species (%)
鹅膏菌科Amanitaceae 2 11 5.26 鹅膏菌属 Amanita 10 4.78
小菇科 Mycenaceae 2 12 5.74 丝盖伞属 Inocybe 5 2.39
多孔菌科 Polyporaceae 8 14 6.70 蜡蘑属 Laccaria 5 2.39
红菇科 Russulaceae 3 23 11.00 小皮伞属 Marasmius 6 2.87
小菇属 Mycena 11 5.26
光柄菇属 Pluteus 5 2.39
红菇属 Russula 17 8.13
栓菌属 Trametes 5 2.39
关闭全屏