DCT-Net: An effective method to diagnose retinal tears from B-scan ultrasound images
Retinal tears (RTs) are usually detected by B-scan ultrasound images, particularly for individuals with complex eye conditions. However, traditional manual techniques for reading ultrasound images have the potential to overlook or inaccurately diagnose conditions. Thus, the development of rapid and...
Saved in:
Published in | Mathematical biosciences and engineering : MBE Vol. 21; no. 1; pp. 1110 - 1124 |
---|---|
Main Authors | , , , , , , , , |
Format | Journal Article |
Language | English |
Published |
United States
AIMS Press
01.01.2024
|
Subjects | |
Online Access | Get full text |
Cover
Loading…
Abstract | Retinal tears (RTs) are usually detected by B-scan ultrasound images, particularly for individuals with complex eye conditions. However, traditional manual techniques for reading ultrasound images have the potential to overlook or inaccurately diagnose conditions. Thus, the development of rapid and accurate approaches for the diagnosis of an RT is highly important and urgent. The present study introduces a novel hybrid deep-learning model called DCT-Net to enable the automatic and precise diagnosis of RTs. The implemented model utilizes a vision transformer as the backbone and feature extractor. Additionally, in order to accommodate the edge characteristics of the lesion areas, a novel module called the residual deformable convolution has been incorporated. Furthermore, normalization is employed to mitigate the issue of overfitting and, a Softmax layer has been included to achieve the final classification following the acquisition of the global and local representations. The study was conducted by using both our proprietary dataset and a publicly available dataset. In addition, interpretability of the trained model was assessed by generating attention maps using the attention rollout approach. On the private dataset, the model demonstrated a high level of performance, with an accuracy of 97.78%, precision of 97.34%, recall rate of 97.13%, and an F1 score of 0.9682. On the other hand, the model developed by using the public funds image dataset demonstrated an accuracy of 83.82%, a sensitivity of 82.69% and a specificity of 82.40%. The findings, therefore present a novel framework for the diagnosis of RTs that is characterized by a high degree of efficiency, accuracy and interpretability. Accordingly, the technology exhibits considerable promise and has the potential to serve as a reliable tool for ophthalmologists. |
---|---|
AbstractList | Retinal tears (RTs) are usually detected by B-scan ultrasound images, particularly for individuals with complex eye conditions. However, traditional manual techniques for reading ultrasound images have the potential to overlook or inaccurately diagnose conditions. Thus, the development of rapid and accurate approaches for the diagnosis of an RT is highly important and urgent. The present study introduces a novel hybrid deep-learning model called DCT-Net to enable the automatic and precise diagnosis of RTs. The implemented model utilizes a vision transformer as the backbone and feature extractor. Additionally, in order to accommodate the edge characteristics of the lesion areas, a novel module called the residual deformable convolution has been incorporated. Furthermore, normalization is employed to mitigate the issue of overfitting and, a Softmax layer has been included to achieve the final classification following the acquisition of the global and local representations. The study was conducted by using both our proprietary dataset and a publicly available dataset. In addition, interpretability of the trained model was assessed by generating attention maps using the attention rollout approach. On the private dataset, the model demonstrated a high level of performance, with an accuracy of 97.78%, precision of 97.34%, recall rate of 97.13%, and an F1 score of 0.9682. On the other hand, the model developed by using the public funds image dataset demonstrated an accuracy of 83.82%, a sensitivity of 82.69% and a specificity of 82.40%. The findings, therefore present a novel framework for the diagnosis of RTs that is characterized by a high degree of efficiency, accuracy and interpretability. Accordingly, the technology exhibits considerable promise and has the potential to serve as a reliable tool for ophthalmologists. |
Author | Li, Wulan Lin, Shishi Liu, Bo Zhu, Qiaolin Li, Ke Zhu, Xixi Ding, Juntao Yan, Wentao Wu, Jianzhang |
Author_xml | – sequence: 1 givenname: Ke surname: Li fullname: Li, Ke organization: The First Affiliated Hospital of Wenzhou Medical University, Wenzhou 325035, China – sequence: 2 givenname: Qiaolin surname: Zhu fullname: Zhu, Qiaolin organization: The Eye Hospital, School of Ophthalmology & Optometry, Wenzhou Medical University, Wenzhou 325027, China – sequence: 3 givenname: Jianzhang surname: Wu fullname: Wu, Jianzhang organization: The Eye Hospital, School of Ophthalmology & Optometry, Wenzhou Medical University, Wenzhou 325027, China – sequence: 4 givenname: Juntao surname: Ding fullname: Ding, Juntao organization: The First Affiliated Hospital of Wenzhou Medical University, Wenzhou 325035, China – sequence: 5 givenname: Bo surname: Liu fullname: Liu, Bo organization: The First Affiliated Hospital of Wenzhou Medical University, Wenzhou 325035, China – sequence: 6 givenname: Xixi surname: Zhu fullname: Zhu, Xixi organization: The Eye Hospital, School of Ophthalmology & Optometry, Wenzhou Medical University, Wenzhou 325027, China – sequence: 7 givenname: Shishi surname: Lin fullname: Lin, Shishi organization: The Eye Hospital, School of Ophthalmology & Optometry, Wenzhou Medical University, Wenzhou 325027, China – sequence: 8 givenname: Wentao surname: Yan fullname: Yan, Wentao organization: The Eye Hospital, School of Ophthalmology & Optometry, Wenzhou Medical University, Wenzhou 325027, China – sequence: 9 givenname: Wulan surname: Li fullname: Li, Wulan organization: The First Affiliated Hospital of Wenzhou Medical University, Wenzhou 325035, China |
BackLink | https://www.ncbi.nlm.nih.gov/pubmed/38303456$$D View this record in MEDLINE/PubMed |
BookMark | eNpNkc9v1DAQhS1URH9x4o58REJpxx47Tri1S6GVKnpZzpZjT5ZUSVxsB4n_ntBdKk7zNPr03uE7ZUdznImxdwIusEV1OXV0IUEqUPUrdiK0FhWAaI7-y8fsNOdHAFSI6g07xgbXrOsTtv282VbfqHziVzOnvidfhl_EJyo_YuAl8jC43Rwz8URlmN3IC7mUeZ_ixK-r7N3Ml7Ekl-MyBz5Mbkf5nL3u3Zjp7eGese9fbrab2-r-4evd5uq-8qihVD44qBuAVhtNSGBM0zRkRO9AiQ4QCbXshApQ16JVTacF1kaBEKYFpwOesbt9b4ju0T6ldT39ttEN9vkR0866VAY_kg3SUB2EUaEJqpZ9F_rQOd0ZjQq8x7Xrw77rKcWfC-VipyF7Gkc3U1yyla2USst1f0U_7lGfYs6J-pdpAfavErsqsQclK_3-ULx0E4UX9p8D_APSJ4Xh |
Cites_doi | 10.1109/CVPR.2016.90 10.1002/ima.22464 10.1016/j.jobe.2022.105246 10.1002/mp.15312 10.1007/978-3-319-24574-4_80 10.1016/j.ajo.2021.02.004 10.1016/S0161-6420(82)34716-8 10.1142/S1793545822500092 10.21037/atm.2019.11.28 10.1038/s41598-018-33013-w 10.1142/S0218488598000094 10.1609/aaai.v34i07.7000 10.1109/CVPR.2016.308 10.1177/14759217211053546 10.1007/s00417-021-05105-3 10.3341/jkos.2014.55.3.387 10.1016/s0161-6420(94)31141-9 10.1109/TMI.2016.2528162 10.1186/s12938-021-00877-5 10.1016/j.ophtha.2008.08.040 10.1016/0002-9394(59)91247-4 |
ContentType | Journal Article |
DBID | CGR CUY CVF ECM EIF NPM AAYXX CITATION 7X8 DOA |
DOI | 10.3934/mbe.2024046 |
DatabaseName | Medline MEDLINE MEDLINE (Ovid) MEDLINE MEDLINE PubMed CrossRef MEDLINE - Academic Directory of Open Access Journals |
DatabaseTitle | MEDLINE Medline Complete MEDLINE with Full Text PubMed MEDLINE (Ovid) CrossRef MEDLINE - Academic |
DatabaseTitleList | MEDLINE - Academic MEDLINE CrossRef |
Database_xml | – sequence: 1 dbid: DOA name: Directory of Open Access Journals url: https://www.doaj.org/ sourceTypes: Open Website – sequence: 2 dbid: NPM name: PubMed url: https://proxy.k.utb.cz/login?url=http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed sourceTypes: Index Database – sequence: 3 dbid: EIF name: MEDLINE url: https://proxy.k.utb.cz/login?url=https://www.webofscience.com/wos/medline/basic-search sourceTypes: Index Database |
DeliveryMethod | fulltext_linktorsrc |
Discipline | Biology |
EISSN | 1551-0018 |
EndPage | 1124 |
ExternalDocumentID | oai_doaj_org_article_d27e6d174d8d462fbdfdba5b75340cc3 10_3934_mbe_2024046 38303456 |
Genre | Journal Article |
GroupedDBID | --- 53G 5GY AENEX ALMA_UNASSIGNED_HOLDINGS CGR CUY CVF EBD EBS ECM EIF EJD EMOBN F5P GROUPED_DOAJ IAO ITC J9A ML0 NPM OK1 P2P RAN SV3 TUS AAYXX CITATION 7X8 |
ID | FETCH-LOGICAL-c350t-cda068009575e3e077888e71fa041b033e352b14d0661948b513674011790a5d3 |
IEDL.DBID | DOA |
ISSN | 1551-0018 |
IngestDate | Thu Jul 04 20:33:49 EDT 2024 Fri Apr 12 04:10:26 EDT 2024 Thu Sep 26 18:20:05 EDT 2024 Sat Sep 28 08:07:32 EDT 2024 |
IsDoiOpenAccess | true |
IsOpenAccess | true |
IsPeerReviewed | true |
IsScholarly | true |
Issue | 1 |
Keywords | automatic diagnosis attention rollout ultrasound image retinal tears vision transformer |
Language | English |
LinkModel | DirectLink |
MergedId | FETCHMERGED-LOGICAL-c350t-cda068009575e3e077888e71fa041b033e352b14d0661948b513674011790a5d3 |
Notes | ObjectType-Article-1 SourceType-Scholarly Journals-1 ObjectType-Feature-2 content type line 23 |
OpenAccessLink | https://doaj.org/article/d27e6d174d8d462fbdfdba5b75340cc3 |
PMID | 38303456 |
PQID | 2922452674 |
PQPubID | 23479 |
PageCount | 15 |
ParticipantIDs | doaj_primary_oai_doaj_org_article_d27e6d174d8d462fbdfdba5b75340cc3 proquest_miscellaneous_2922452674 crossref_primary_10_3934_mbe_2024046 pubmed_primary_38303456 |
PublicationCentury | 2000 |
PublicationDate | 2024-01-01 |
PublicationDateYYYYMMDD | 2024-01-01 |
PublicationDate_xml | – month: 01 year: 2024 text: 2024-01-01 day: 01 |
PublicationDecade | 2020 |
PublicationPlace | United States |
PublicationPlace_xml | – name: United States |
PublicationTitle | Mathematical biosciences and engineering : MBE |
PublicationTitleAlternate | Math Biosci Eng |
PublicationYear | 2024 |
Publisher | AIMS Press |
Publisher_xml | – name: AIMS Press |
References | key-10.3934/mbe.2024046-30 key-10.3934/mbe.2024046-10 key-10.3934/mbe.2024046-32 key-10.3934/mbe.2024046-31 key-10.3934/mbe.2024046-23 key-10.3934/mbe.2024046-22 key-10.3934/mbe.2024046-25 key-10.3934/mbe.2024046-24 key-10.3934/mbe.2024046-27 key-10.3934/mbe.2024046-26 key-10.3934/mbe.2024046-29 key-10.3934/mbe.2024046-28 key-10.3934/mbe.2024046-7 key-10.3934/mbe.2024046-6 key-10.3934/mbe.2024046-5 key-10.3934/mbe.2024046-4 key-10.3934/mbe.2024046-3 key-10.3934/mbe.2024046-2 key-10.3934/mbe.2024046-1 key-10.3934/mbe.2024046-9 key-10.3934/mbe.2024046-21 key-10.3934/mbe.2024046-8 key-10.3934/mbe.2024046-20 key-10.3934/mbe.2024046-12 key-10.3934/mbe.2024046-34 key-10.3934/mbe.2024046-11 key-10.3934/mbe.2024046-33 key-10.3934/mbe.2024046-14 key-10.3934/mbe.2024046-36 key-10.3934/mbe.2024046-13 key-10.3934/mbe.2024046-35 key-10.3934/mbe.2024046-16 key-10.3934/mbe.2024046-38 key-10.3934/mbe.2024046-15 key-10.3934/mbe.2024046-37 key-10.3934/mbe.2024046-18 key-10.3934/mbe.2024046-17 key-10.3934/mbe.2024046-39 key-10.3934/mbe.2024046-19 |
References_xml | – ident: key-10.3934/mbe.2024046-15 – ident: key-10.3934/mbe.2024046-19 doi: 10.1109/CVPR.2016.90 – ident: key-10.3934/mbe.2024046-11 – ident: key-10.3934/mbe.2024046-38 doi: 10.1002/ima.22464 – ident: key-10.3934/mbe.2024046-17 – ident: key-10.3934/mbe.2024046-37 doi: 10.1016/j.jobe.2022.105246 – ident: key-10.3934/mbe.2024046-27 – ident: key-10.3934/mbe.2024046-13 doi: 10.1002/mp.15312 – ident: key-10.3934/mbe.2024046-34 doi: 10.1007/978-3-319-24574-4_80 – ident: key-10.3934/mbe.2024046-7 doi: 10.1016/j.ajo.2021.02.004 – ident: key-10.3934/mbe.2024046-5 doi: 10.1016/S0161-6420(82)34716-8 – ident: key-10.3934/mbe.2024046-23 – ident: key-10.3934/mbe.2024046-21 – ident: key-10.3934/mbe.2024046-25 – ident: key-10.3934/mbe.2024046-12 doi: 10.1142/S1793545822500092 – ident: key-10.3934/mbe.2024046-29 – ident: key-10.3934/mbe.2024046-14 – ident: key-10.3934/mbe.2024046-8 doi: 10.21037/atm.2019.11.28 – ident: key-10.3934/mbe.2024046-35 doi: 10.1038/s41598-018-33013-w – ident: key-10.3934/mbe.2024046-20 doi: 10.1142/S0218488598000094 – ident: key-10.3934/mbe.2024046-10 – ident: key-10.3934/mbe.2024046-16 doi: 10.1609/aaai.v34i07.7000 – ident: key-10.3934/mbe.2024046-30 doi: 10.1109/CVPR.2016.308 – ident: key-10.3934/mbe.2024046-36 doi: 10.1177/14759217211053546 – ident: key-10.3934/mbe.2024046-18 – ident: key-10.3934/mbe.2024046-31 – ident: key-10.3934/mbe.2024046-9 doi: 10.1007/s00417-021-05105-3 – ident: key-10.3934/mbe.2024046-4 doi: 10.3341/jkos.2014.55.3.387 – ident: key-10.3934/mbe.2024046-1 doi: 10.1016/s0161-6420(94)31141-9 – ident: key-10.3934/mbe.2024046-6 doi: 10.1109/TMI.2016.2528162 – ident: key-10.3934/mbe.2024046-39 doi: 10.1142/S1793545822500092 – ident: key-10.3934/mbe.2024046-22 – ident: key-10.3934/mbe.2024046-33 doi: 10.1186/s12938-021-00877-5 – ident: key-10.3934/mbe.2024046-2 doi: 10.1016/j.ophtha.2008.08.040 – ident: key-10.3934/mbe.2024046-24 – ident: key-10.3934/mbe.2024046-3 doi: 10.1016/0002-9394(59)91247-4 – ident: key-10.3934/mbe.2024046-32 doi: 10.1016/j.ophtha.2008.08.040 – ident: key-10.3934/mbe.2024046-26 – ident: key-10.3934/mbe.2024046-28 |
SSID | ssj0034334 |
Score | 2.3488438 |
Snippet | Retinal tears (RTs) are usually detected by B-scan ultrasound images, particularly for individuals with complex eye conditions. However, traditional manual... |
SourceID | doaj proquest crossref pubmed |
SourceType | Open Website Aggregation Database Index Database |
StartPage | 1110 |
SubjectTerms | attention rollout automatic diagnosis Electric Power Supplies Humans Neuroimaging Retinal Perforations retinal tears Ultrasonography ultrasound image vision transformer |
Title | DCT-Net: An effective method to diagnose retinal tears from B-scan ultrasound images |
URI | https://www.ncbi.nlm.nih.gov/pubmed/38303456 https://search.proquest.com/docview/2922452674 https://doaj.org/article/d27e6d174d8d462fbdfdba5b75340cc3 |
Volume | 21 |
hasFullText | 1 |
inHoldings | 1 |
isFullTextHit | |
isPrint | |
link | http://utb.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwrV09a8MwEBWlUOhS-t30CxWymtjWyR_dkrQhFJopgWxGiiQoNE6JnaH_vneWE7KULl2NsaX3ZOmddXrHWNdl2sk8QgZSCQGAcEGW5RAYcAqcSXOnKVB8nyTjGbzN5Xyv1BflhHl7YA9cz8SpTQzqZpMZSGKnjTNaSY0yG8LFwvt8RnIbTPk5WIAQ4E_jiVxAb6nJERMXL9K5e-tPY9P_u7Zs1pjRKTtpxSHv-0adsQNbnrMjXy7y-4JNX4bTYGLrZ94vuU_EwLmK-yLQvF5x4_PmLKejifQkStatOB0h4YOgQhT55rNeq4pqKfGPJc4l1SWbjV6nw3HQVkUIFkKGdbAwiuploDRKpRU2TCmItWnkVAiRDoWwqKl0BAbFRJRDpiW5skHj_RYqacQVOyxXpb1hHKSTjf-OSi1kKtYav-9EOpeRLb3WHdbdYlV8efOLAoMGgrRASIsW0g4bEI67W8ixurmAPBYtj8VfPHbY05aFAkc4bVuo0q42VRHnMW0PYx867NrTs3sVxtdItkxu_6MJd-yYeuR_sNyzw3q9sQ8oOWr92IyuH49m0ns |
link.rule.ids | 315,786,790,870,2115,27957,27958 |
linkProvider | Directory of Open Access Journals |
openUrl | ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=DCT-Net%3A+An+effective+method+to+diagnose+retinal+tears+from+B-scan+ultrasound+images&rft.jtitle=Mathematical+biosciences+and+engineering+%3A+MBE&rft.au=Li%2C+Ke&rft.au=Zhu%2C+Qiaolin&rft.au=Wu%2C+Jianzhang&rft.au=Ding%2C+Juntao&rft.date=2024-01-01&rft.eissn=1551-0018&rft.volume=21&rft.issue=1&rft.spage=1110&rft.epage=1124&rft_id=info:doi/10.3934%2Fmbe.2024046&rft.externalDBID=NO_FULL_TEXT |
thumbnail_l | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=1551-0018&client=summon |
thumbnail_m | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=1551-0018&client=summon |
thumbnail_s | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=1551-0018&client=summon |