Self-supervised Deep-Learning Segmentation of Corneal Endothelium Specular Microscopy Images
Computerized medical evaluation of the corneal endothelium is challenging because it requires costly equipment and specialized personnel, not to mention that conventional techniques require manual annotations that are difficult to acquire. This study aims to obtain reliable segmentations without req...
- Autores:
-
Sanchez, Sergio
Quintero, Fernando
Mendoza, Kevin
Prada, Prada
Tello, Alejandro
Galvis, Virgilio
Romero, Lenny A
Marrugo, Andres G
- Tipo de recurso:
- Fecha de publicación:
- 2023
- Institución:
- Universidad Tecnológica de Bolívar
- Repositorio:
- Repositorio Institucional UTB
- Idioma:
- eng
- OAI Identifier:
- oai:repositorio.utb.edu.co:20.500.12585/12633
- Acceso en línea:
- https://hdl.handle.net/20.500.12585/12633
- Palabra clave:
- Training
Semantic segmentation
Microscopy
Self-supervised learning
Data models
Personnel
LEMB
- Rights
- openAccess
- License
- http://creativecommons.org/licenses/by-nc-sa/4.0/
| id |
UTB2_14cc53d8591d060b31ce9db7549a0f07 |
|---|---|
| oai_identifier_str |
oai:repositorio.utb.edu.co:20.500.12585/12633 |
| network_acronym_str |
UTB2 |
| network_name_str |
Repositorio Institucional UTB |
| repository_id_str |
|
| dc.title.spa.fl_str_mv |
Self-supervised Deep-Learning Segmentation of Corneal Endothelium Specular Microscopy Images |
| title |
Self-supervised Deep-Learning Segmentation of Corneal Endothelium Specular Microscopy Images |
| spellingShingle |
Self-supervised Deep-Learning Segmentation of Corneal Endothelium Specular Microscopy Images Training Semantic segmentation Microscopy Self-supervised learning Data models Personnel LEMB |
| title_short |
Self-supervised Deep-Learning Segmentation of Corneal Endothelium Specular Microscopy Images |
| title_full |
Self-supervised Deep-Learning Segmentation of Corneal Endothelium Specular Microscopy Images |
| title_fullStr |
Self-supervised Deep-Learning Segmentation of Corneal Endothelium Specular Microscopy Images |
| title_full_unstemmed |
Self-supervised Deep-Learning Segmentation of Corneal Endothelium Specular Microscopy Images |
| title_sort |
Self-supervised Deep-Learning Segmentation of Corneal Endothelium Specular Microscopy Images |
| dc.creator.fl_str_mv |
Sanchez, Sergio Quintero, Fernando Mendoza, Kevin Prada, Prada Tello, Alejandro Galvis, Virgilio Romero, Lenny A Marrugo, Andres G |
| dc.contributor.author.none.fl_str_mv |
Sanchez, Sergio Quintero, Fernando Mendoza, Kevin Prada, Prada Tello, Alejandro Galvis, Virgilio Romero, Lenny A Marrugo, Andres G |
| dc.contributor.other.none.fl_str_mv |
Mendoza, Kevin |
| dc.subject.keywords.spa.fl_str_mv |
Training Semantic segmentation Microscopy Self-supervised learning Data models Personnel |
| topic |
Training Semantic segmentation Microscopy Self-supervised learning Data models Personnel LEMB |
| dc.subject.armarc.none.fl_str_mv |
LEMB |
| description |
Computerized medical evaluation of the corneal endothelium is challenging because it requires costly equipment and specialized personnel, not to mention that conventional techniques require manual annotations that are difficult to acquire. This study aims to obtain reliable segmentations without requiring large data sets labeled by expert personnel. To address this problem, we use the Barlow Twins approach to pre-train the encoder of a UNet model in an unsupervised manner. Then, with few labeled data, we train the segmentation. Encouraging results show that it is possible to address the challenge of limited data availability using self-supervised learning. This model achieved a precision of 86\%, obtaining a satisfactory performance. Using many images to learn good representations and a few labeled images to learn the semantic segmentation task is feasible. |
| publishDate |
2023 |
| dc.date.issued.none.fl_str_mv |
2023 |
| dc.date.accessioned.none.fl_str_mv |
2024-02-12T19:17:55Z |
| dc.date.available.none.fl_str_mv |
2024-02-12T19:17:55Z |
| dc.date.submitted.none.fl_str_mv |
2024 |
| dc.type.coarversion.fl_str_mv |
http://purl.org/coar/version/c_b1a7d7d4d402bcce |
| dc.type.coar.fl_str_mv |
http://purl.org/coar/resource_type/c_2df8fbb1 |
| dc.type.driver.spa.fl_str_mv |
info:eu-repo/semantics/article |
| dc.type.hasversion.spa.fl_str_mv |
info:eu-repo/semantics/draft |
| dc.type.spa.spa.fl_str_mv |
http://purl.org/coar/resource_type/c_6501 |
| status_str |
draft |
| dc.identifier.citation.spa.fl_str_mv |
Sánchez, S., Mendoza, K., Quintero, F. J., Prada, A. M., Tello, A., Galvis, V., Romero, L. A., & Marrugo, A. G. (2023). Self-supervised deep-learning segmentation of corneal endothelium specular microscopy images. 2023 IEEE Colombian Conference on Applications of Computational Intelligence (ColCACI), 1–5. |
| dc.identifier.uri.none.fl_str_mv |
https://hdl.handle.net/20.500.12585/12633 |
| dc.identifier.doi.none.fl_str_mv |
DOI: 10.1109/ColCACI59285.2023.10226148 |
| dc.identifier.instname.spa.fl_str_mv |
Universidad Tecnológica de Bolívar |
| dc.identifier.reponame.spa.fl_str_mv |
Repositorio Universidad Tecnológica de Bolívar |
| identifier_str_mv |
Sánchez, S., Mendoza, K., Quintero, F. J., Prada, A. M., Tello, A., Galvis, V., Romero, L. A., & Marrugo, A. G. (2023). Self-supervised deep-learning segmentation of corneal endothelium specular microscopy images. 2023 IEEE Colombian Conference on Applications of Computational Intelligence (ColCACI), 1–5. DOI: 10.1109/ColCACI59285.2023.10226148 Universidad Tecnológica de Bolívar Repositorio Universidad Tecnológica de Bolívar |
| url |
https://hdl.handle.net/20.500.12585/12633 |
| dc.language.iso.spa.fl_str_mv |
eng |
| language |
eng |
| dc.rights.coar.fl_str_mv |
http://purl.org/coar/access_right/c_abf2 |
| dc.rights.uri.*.fl_str_mv |
http://creativecommons.org/licenses/by-nc-sa/4.0/ |
| dc.rights.accessrights.spa.fl_str_mv |
info:eu-repo/semantics/openAccess |
| dc.rights.cc.*.fl_str_mv |
Atribución-NoComercial-CompartirIgual 4.0 Internacional |
| rights_invalid_str_mv |
http://creativecommons.org/licenses/by-nc-sa/4.0/ Atribución-NoComercial-CompartirIgual 4.0 Internacional http://purl.org/coar/access_right/c_abf2 |
| eu_rights_str_mv |
openAccess |
| dc.format.extent.none.fl_str_mv |
14 páginas |
| dc.format.mimetype.spa.fl_str_mv |
application/pdf |
| dc.publisher.place.spa.fl_str_mv |
Cartagena de Indias |
| dc.source.spa.fl_str_mv |
2023 IEEE Colombian Conference on Applications of Computational Intelligence (ColCACI) |
| institution |
Universidad Tecnológica de Bolívar |
| bitstream.url.fl_str_mv |
https://repositorio.utb.edu.co/bitstream/20.500.12585/12633/1/SpringerCOLCACI.pdf https://repositorio.utb.edu.co/bitstream/20.500.12585/12633/2/license_rdf https://repositorio.utb.edu.co/bitstream/20.500.12585/12633/3/license.txt https://repositorio.utb.edu.co/bitstream/20.500.12585/12633/4/SpringerCOLCACI.pdf.txt https://repositorio.utb.edu.co/bitstream/20.500.12585/12633/5/SpringerCOLCACI.pdf.jpg |
| bitstream.checksum.fl_str_mv |
7bf7ee2280a7116e77863a4e4253f9ee 934f4ca17e109e0a05eaeaba504d7ce4 e20ad307a1c5f3f25af9304a7a7c86b6 30dc708a0601f126168d0e06ce28f755 0f4ff1debde5bc066e415a6b305c1139 |
| bitstream.checksumAlgorithm.fl_str_mv |
MD5 MD5 MD5 MD5 MD5 |
| repository.name.fl_str_mv |
Repositorio Institucional UTB |
| repository.mail.fl_str_mv |
repositorioutb@utb.edu.co |
| _version_ |
1814021614887501824 |
| spelling |
Sanchez, Sergio6600d51e-4c2e-4d6c-9bc8-2713e182f4faQuintero, Fernandof07ea69c-0c88-4255-be83-d0aa76f66768Mendoza, Kevin135a7756-119e-4c7c-ae0b-d7111a8a3a61Prada, Prada2fbf01bb-7daf-4db0-8b21-82f4352fb43cTello, Alejandrob88c245a-e5d9-4feb-a8e5-fc2a6555415dGalvis, Virgilio85e1c5d8-b4a4-4bed-828d-267cd8ca4b5bRomero, Lenny A4e34aa8a-f981-4e1d-ae32-d45acb6abcf9Marrugo, Andres G3d6cd388-d48f-4669-934f-49ca4179f542Mendoza, Kevin2024-02-12T19:17:55Z2024-02-12T19:17:55Z20232024Sánchez, S., Mendoza, K., Quintero, F. J., Prada, A. M., Tello, A., Galvis, V., Romero, L. A., & Marrugo, A. G. (2023). Self-supervised deep-learning segmentation of corneal endothelium specular microscopy images. 2023 IEEE Colombian Conference on Applications of Computational Intelligence (ColCACI), 1–5.https://hdl.handle.net/20.500.12585/12633DOI: 10.1109/ColCACI59285.2023.10226148Universidad Tecnológica de BolívarRepositorio Universidad Tecnológica de BolívarComputerized medical evaluation of the corneal endothelium is challenging because it requires costly equipment and specialized personnel, not to mention that conventional techniques require manual annotations that are difficult to acquire. This study aims to obtain reliable segmentations without requiring large data sets labeled by expert personnel. To address this problem, we use the Barlow Twins approach to pre-train the encoder of a UNet model in an unsupervised manner. Then, with few labeled data, we train the segmentation. Encouraging results show that it is possible to address the challenge of limited data availability using self-supervised learning. This model achieved a precision of 86\%, obtaining a satisfactory performance. Using many images to learn good representations and a few labeled images to learn the semantic segmentation task is feasible.Minciencias14 páginasapplication/pdfenghttp://creativecommons.org/licenses/by-nc-sa/4.0/info:eu-repo/semantics/openAccessAtribución-NoComercial-CompartirIgual 4.0 Internacionalhttp://purl.org/coar/access_right/c_abf22023 IEEE Colombian Conference on Applications of Computational Intelligence (ColCACI)Self-supervised Deep-Learning Segmentation of Corneal Endothelium Specular Microscopy Imagesinfo:eu-repo/semantics/articleinfo:eu-repo/semantics/drafthttp://purl.org/coar/resource_type/c_6501http://purl.org/coar/version/c_b1a7d7d4d402bccehttp://purl.org/coar/resource_type/c_2df8fbb1TrainingSemantic segmentationMicroscopySelf-supervised learningData modelsPersonnelLEMBCartagena de IndiasInvestigadoresJeang, L.J., Margo, C.E., Espana, E.M.: Diseases of the corneal endothelium. Experimental eye research 205, 108495 (2021)Catala, P., Thuret, G., Skottman, H., Mehta, J.S., Parekh, M., Dhubhghaill, S.N., Collin, R.W., Nuijts, R.M., Ferrari, S., LaPointe, V.L., et al.: Approaches for corneal endothelium regenerative medicine. Progress in retinal and eye research 87, 100987 (2022)Sierra, J.S., Pineda, J., Rueda, D., Tello, A., Prada, A.M., Galvis, V., Volpe, G., Millan, M.S., Romero, L.A., Marrugo, A.G.: Corneal endothelium assessment in specular microscopy images with fuchs’ dystrophy via deep regression of signed distance maps. Biomedical optics express 14(1), 335–351 (2023)Knauer, C., Pfeiffer, N.: The value of vision. Graefe’s Archive for Clinical and Experimental Ophthalmology 246, 477–482 (2008)Huang, J., Maram, J., Tepelus, T.C., Sadda, S.R., Chopra, V., Lee, O.L.: Com parison of noncontact specular and confocal microscopy for evaluation of corneal endothelium. Eye & Contact Lens: Science & Clinical Practice 44, 144–150 (2017)Price, M.O., Fairchild, K.M., Price, F.W.: Comparison of manual and automated endothelial cell density analysis in normal eyes and dsek eyes. Cornea 32(5), 567–573 (2013) https://doi.org/10.1097/ico.0b013e31825de8faLuft, N., Hirnschall, N., Schuschitz, S., Draschl, P., Findl, O.: Compari son of 4 specular microscopes in healthy eyes and eyes with cornea guttata or corneal grafts. Cornea 34(4), 381–386 (2015) https://doi.org/10.1097/ico. 0000000000000385Gasser, L., Reinhard, T., B¨ohringer, D.: Comparison of corneal endothelial cell measurements by two non-contact specular microscopes. BMC ophthalmology 15, 87 (2015) https://doi.org/10.1186/s12886-015-0068-1Selig, B., Vermeer, K., Rieger, B., Hillenaar, T., Luengo Hendriks, C.: Fully auto matic evaluation of the corneal endothelium from in vivo confocal microscopy. BMC medical imaging 15, 13 (2015) https://doi.org/10.1186/s12880-015-0054-3Shilpashree, P., Kaggere, S., Sudhir, R., Srinivas, S.: Automated image segmen tation of the corneal endothelium in patients with fuchs dystrophy. Translational Vision Science & Technology 10, 27 (2021) https://doi.org/10.1167/tvst.10.13.27Daniel, M., Atzrodt, L., Bucher, F., Wacker, K., B¨ohringer, S., Reinhard, T., B¨ohringer, D.: Automated segmentation of the corneal endothelium in a large set of ‘real-world’ specular microscopy images using the u-net architecture. Scientific Reports 9 (2019) https://doi.org/10.1038/s41598-019-41034-2Vigueras-Guill´en, J., Rooij, J., Dooren, B., Lemij, H., Islamaj, E., Van Vliet, L., Vermeer, K.: Denseunets with feedback non-local attention for the segmentation of specular microscopy images of the corneal endothelium with guttae. Scientific Reports 12 (2022) https://doi.org/10.1038/s41598-022-18180-1Caron, M., Touvron, H., Misra, I., J´egou, H., Mairal, J., Bojanowski, P., Joulin, A.: Emerging Properties in Self-Supervised Vision Transformers (2021)Zbontar, J., Jing, L., Misra, I., LeCun, Y., Deny, S.: Barlow twins: Self-supervised learning via redundancy reduction. In: International Conference on Machine Learning, pp. 12310–12320 (2021). PMLRPunn, N.S., Agarwal, S.: Bt-unet: A self-supervised learning framework for biomedical image segmentation using barlow twins with u-net models. Machine Learning 111(12), 4585–4600 (2022)Jiao, R., Zhang, Y., Ding, L., Cai, R., Zhang, J.: Learning with Limited Annotations: A Survey on Deep Semi-Supervised Learning for Medical Image Segmentation (2022)Balestriero, R., LeCun, Y.: Contrastive and Non-Contrastive Self-Supervised Learning Recover Global and Local Spectral Embedding Methods (2022)Grill, J.-B., Strub, F., Altch´e, F., Tallec, C., Richemond, P.H., Buchatskaya, E., Doersch, C., Pires, B.A., Guo, Z.D., Azar, M.G., Piot, B., Kavukcuoglu, K., Munos, R., Valko, M.: Bootstrap your own latent: A new approach to self-supervised Learning (2020)Liu, C., Amodio, M., Shen, L.L., Gao, F., Avesta, A., Aneja, S., Wang, J.C., Priore, L.V.D., Krishnaswamy, S.: CUTS: A Fully Unsupervised Framework for Medical Image Segmentation (2023Felfeliyan, B., Hareendranathan, A., Kuntze, G., Cornell, D., Forkert, N.D., Jaremko, J.L., Ronsky, J.L.: Self-Supervised-RCNN for Medical Image Segmen tation with Limited Data Annotation (2022)Ronneberger, O., Fischer, P., Brox, T.: U-net: Convolutional networks for biomed ical image segmentation. In: Medical Image Computing and Computer-Assisted Intervention–MICCAI 2015: 18th International Conference, Munich, Germany, October 5-9, 2015, Proceedings, Part III 18, pp. 234–241 (2015). SpringerMarsocci, V., Scardapane, S.: Continual barlow twins: Continual self-supervised learning for remote sensing semantic segmentation. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing 16, 5049–5060 (2023) https://doi.org/10.1109/JSTARS.2023.3280029http://purl.org/coar/resource_type/c_2df8fbb1ORIGINALSpringerCOLCACI.pdfSpringerCOLCACI.pdfapplication/pdf851136https://repositorio.utb.edu.co/bitstream/20.500.12585/12633/1/SpringerCOLCACI.pdf7bf7ee2280a7116e77863a4e4253f9eeMD51CC-LICENSElicense_rdflicense_rdfapplication/rdf+xml; charset=utf-81031https://repositorio.utb.edu.co/bitstream/20.500.12585/12633/2/license_rdf934f4ca17e109e0a05eaeaba504d7ce4MD52LICENSElicense.txtlicense.txttext/plain; charset=utf-83182https://repositorio.utb.edu.co/bitstream/20.500.12585/12633/3/license.txte20ad307a1c5f3f25af9304a7a7c86b6MD53TEXTSpringerCOLCACI.pdf.txtSpringerCOLCACI.pdf.txtExtracted texttext/plain29643https://repositorio.utb.edu.co/bitstream/20.500.12585/12633/4/SpringerCOLCACI.pdf.txt30dc708a0601f126168d0e06ce28f755MD54THUMBNAILSpringerCOLCACI.pdf.jpgSpringerCOLCACI.pdf.jpgGenerated Thumbnailimage/jpeg4126https://repositorio.utb.edu.co/bitstream/20.500.12585/12633/5/SpringerCOLCACI.pdf.jpg0f4ff1debde5bc066e415a6b305c1139MD5520.500.12585/12633oai:repositorio.utb.edu.co:20.500.12585/126332024-02-13 00:16:53.286Repositorio Institucional UTBrepositorioutb@utb.edu.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 |