{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,3,18]],"date-time":"2026-03-18T13:45:28Z","timestamp":1773841528772,"version":"3.50.1"},"reference-count":45,"publisher":"Springer Science and Business Media LLC","issue":"5","license":[{"start":{"date-parts":[[2021,6,16]],"date-time":"2021-06-16T00:00:00Z","timestamp":1623801600000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2021,6,16]],"date-time":"2021-06-16T00:00:00Z","timestamp":1623801600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["Cogn Comput"],"published-print":{"date-parts":[[2022,9]]},"abstract":"<jats:title>Abstract\n<\/jats:title><jats:p>Clinical keratoconus (KCN) detection is a challenging and time-consuming task. In the diagnosis process, ophthalmologists must revise demographic and clinical ophthalmic examinations. The latter include slit-lamb, corneal topographic maps, and Pentacam indices (PI). We propose an Ensemble of Deep Transfer Learning (EDTL) based on corneal topographic maps. We consider four pretrained networks, SqueezeNet (SqN), AlexNet (AN), ShuffleNet (SfN), and MobileNet-v2 (MN), and fine-tune them on a dataset of KCN and normal cases, each including four topographic maps. We also consider a PI classifier. Then, our EDTL method combines the output probabilities of each of the five classifiers to obtain a decision based on the fusion of probabilities. Individually, the classifier based on PI achieved 93.1% accuracy, whereas the deep classifiers reached classification accuracies over 90% only in isolated cases. Overall, the average accuracy of the deep networks over the four corneal maps ranged from 86% (SfN) to 89.9% (AN). The classifier ensemble increased the accuracy of the deep classifiers based on corneal maps to values ranging (92.2% to 93.1%) for SqN and (93.1% to 94.8%) for AN. Including in the ensemble-specific combinations of corneal maps\u2019 classifiers and PI increased the accuracy to 98.3%. Moreover, visualization of first learner filters in the networks and Grad-CAMs confirmed that the networks had learned relevant clinical features. This study shows the potential of creating ensembles of deep classifiers fine-tuned with a transfer learning strategy as it resulted in an improved accuracy while showing learnable filters and Grad-CAMs that agree with clinical knowledge. This is a step further towards the potential clinical deployment of an improved computer-assisted diagnosis system for KCN detection to help ophthalmologists to confirm the clinical decision and to perform fast and accurate KCN treatment.<\/jats:p>","DOI":"10.1007\/s12559-021-09880-3","type":"journal-article","created":{"date-parts":[[2021,6,16]],"date-time":"2021-06-16T10:03:21Z","timestamp":1623837801000},"page":"1627-1642","update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":36,"title":["Deep Transfer Learning for Improved Detection of Keratoconus using Corneal Topographic Maps"],"prefix":"10.1007","volume":"14","author":[{"ORCID":"https:\/\/orcid.org\/0000-0003-2738-8896","authenticated-orcid":false,"given":"Ali H.","family":"Al-Timemy","sequence":"first","affiliation":[]},{"given":"Nebras H.","family":"Ghaeb","sequence":"additional","affiliation":[]},{"given":"Zahraa M.","family":"Mosa","sequence":"additional","affiliation":[]},{"ORCID":"https:\/\/orcid.org\/0000-0002-2105-8725","authenticated-orcid":false,"given":"Javier","family":"Escudero","sequence":"additional","affiliation":[]}],"member":"297","published-online":{"date-parts":[[2021,6,16]]},"reference":[{"key":"9880_CR1","doi-asserted-by":"crossref","unstructured":"Yousefi S, Yousefi E, Takahashi H, Hayashi T, Tampo H, Inoda S, et al. Keratoconus severity identification using unsupervised machine learning. PLoS One. Public Library of Science San Francisco, CA USA; 2018;13:e0205998.","DOI":"10.1371\/journal.pone.0205998"},{"key":"9880_CR2","doi-asserted-by":"publisher","first-page":"151","DOI":"10.1016\/S1532-0464(02)00513-0","volume":"35","author":"PA Accardo","year":"2002","unstructured":"Accardo PA, Pensiero S. Neural network-based system for early keratoconus detection from corneal topography. J Biomed Inform Elsevier. 2002;35:151\u20139.","journal-title":"J Biomed Inform Elsevier"},{"key":"9880_CR3","doi-asserted-by":"crossref","unstructured":"Belin MW, Ambr\u00f3sio Jr R. Scheimpflug imaging for keratoconus and ectatic disease. Indian J Ophthalmol. Wolters Kluwer--Medknow Publications; 2013;61:401.","DOI":"10.4103\/0301-4738.116059"},{"key":"9880_CR4","doi-asserted-by":"crossref","unstructured":"Castillo JH, Hanna R, Berkowitz E, Tiosano B. Wavefront analysis for keratoconus. Int J Keratoconus Ectatic Corneal Dis. Jaypee Brothers Medical Publishers Ltd.; 2014;3:76.","DOI":"10.5005\/jp-journals-10025-1083"},{"key":"9880_CR5","doi-asserted-by":"crossref","unstructured":"Karamichos D, Hjortdal J. Keratoconus: tissue engineering and biomaterials. J Funct Biomater. Multidisciplinary Digital Publishing Institute; 2014;5:111\u201334.","DOI":"10.3390\/jfb5030111"},{"key":"9880_CR6","doi-asserted-by":"crossref","unstructured":"Gatinel D, Malet J, Hoang-Xuan T, Azar DT. Corneal elevation topography: best fit sphere, elevation distance, asphericity, toricity and clinical implications. Cornea. NIH Public Access; 2011;30:508.","DOI":"10.1097\/ICO.0b013e3181fb4fa7"},{"key":"9880_CR7","unstructured":"Smolek MK, Klyce SD. Current keratoconus detection methods compared with a neural network approach. Invest Ophthalmol Vis Sci. The Association for Research in Vision and Ophthalmology; 1997;38:2290\u20139."},{"key":"9880_CR8","doi-asserted-by":"crossref","unstructured":"Aslani F, Khorrami-Nejad M, Amiri MA, Hashemian H, Askarizadeh F, Khosravi B. Characteristics of posterior corneal astigmatism in different stages of keratoconus. J Ophthalmic Vis Res. Wolters Kluwer--Medknow Publications; 2018;13:3.","DOI":"10.4103\/jovr.jovr_217_16"},{"key":"9880_CR9","first-page":"1","volume":"2","author":"TK Yoo","year":"2019","unstructured":"Yoo TK, Ryu IH, Lee G, Kim Y, Kim JK, Lee IS, et al. Adopting machine learning to automatically identify candidate patients for corneal refractive surgery. NPJ Digit Med Nature Publishing Group. 2019;2:1\u20139.","journal-title":"NPJ Digit Med Nature Publishing Group"},{"key":"9880_CR10","doi-asserted-by":"publisher","unstructured":"Issarti I, Consejo A, Koppen C, Rozema JJ, Jim\u00e9nez-garc\u00eda M, Hershko S. Computer aided diagnosis for suspect keratoconus detection. Comput Biol Med [Internet]. Elsevier Ltd; 2019;109:33\u201342. Available from:\u00a0https:\/\/doi.org\/10.1016\/j.compbiomed.2019.04.024","DOI":"10.1016\/j.compbiomed.2019.04.024"},{"key":"9880_CR11","first-page":"4","volume":"16","author":"ZM Mosa","year":"2019","unstructured":"Mosa ZM, Ghaeb NH, Ali AH. Detecting keratoconus by using SVM and decision tree classifiers with the aid of image processing. Baghdad Sci J. 2019;16:4.","journal-title":"Baghdad Sci J"},{"key":"9880_CR12","unstructured":"Sinjab MM. Step by Step\u00ae: reading Pentacam topography: basics and case study series. Jaypee Brothers Medical Publishers; 2015."},{"key":"9880_CR13","unstructured":"Maeda N, Klyce SD, Smolek MK. Neural network classification of corneal topography preliminary demonstration. Invest Ophthalmol Vis Sci. 1995;36."},{"key":"9880_CR14","doi-asserted-by":"publisher","first-page":"4","DOI":"10.1109\/JBHI.2016.2636665","volume":"21","author":"D Rav","year":"2017","unstructured":"Rav D, Wong C, Deligianni F, Berthelot M, Andreu-perez J, Lo B. Deep learning for health informatics. IEEE J Biomed Heal Informatics. 2017;21:4\u201321.","journal-title":"IEEE J Biomed Heal Informatics"},{"key":"9880_CR15","doi-asserted-by":"crossref","unstructured":"Alzubaidi L, Fadhel MA, Oleiwi SR, Al-Shamma O, Zhang J. DFU_QUTNet: diabetic foot ulcer classification using novel deep convolutional neural network. Multimed Tools Appl. Springer; 2019;1\u201323.","DOI":"10.1007\/s11042-019-07820-w"},{"key":"9880_CR16","doi-asserted-by":"publisher","first-page":"1627","DOI":"10.1109\/JBHI.2015.2425041","volume":"19","author":"H Chen","year":"2015","unstructured":"Chen H, Ni D, Qin J, Li S, Yang X, Wang T, et al. Standard plane localization in fetal ultrasound via domain transferred deep neural networks. IEEE J Biomed Heal informatics IEEE. 2015;19:1627\u201336.","journal-title":"IEEE J Biomed Heal informatics IEEE"},{"key":"9880_CR17","first-page":"1","volume":"2018","author":"M Norouzifard","year":"2018","unstructured":"Norouzifard M, Nemati A, GholamHosseini H, Klette R, Nouri-Mahdavi K, Automated YS, deeptransfer learning: proposal of a system for clinical testing. GDU. Int Conf Image Vis Comput New Zeal. IEEE. 2018;2018:1\u20136.","journal-title":"IEEE"},{"key":"9880_CR18","doi-asserted-by":"publisher","first-page":"204","DOI":"10.1007\/s40135-019-00218-9","volume":"7","author":"BT Lopes","year":"2019","unstructured":"Lopes BT, Eliasy A, Ambrosio R Jr. Artificial intelligence in corneal diagnosis: where are we? Curr Ophthalmol Rep Current Ophthalmology Reports. 2019;7:204\u201311.","journal-title":"Curr Ophthalmol Rep Current Ophthalmology Reports"},{"key":"9880_CR19","doi-asserted-by":"publisher","first-page":"167","DOI":"10.1136\/bjophthalmol-2018-313173","volume":"103","author":"D Shu","year":"2019","unstructured":"Shu D, Ting W, Pasquale LR, Peng L, Campbell JP, Lee AY, et al. Artificial intelligence and deep learning in ophthalmology. Br J Ophthalmol. 2019;103:167\u201375.","journal-title":"Br J Ophthalmol"},{"key":"9880_CR20","doi-asserted-by":"crossref","unstructured":"Hallak JA, Azar DT. The AI revolution and how to prepare for it. Transl Vis Sci Technol. The Association for Research in Vision and Ophthalmology; 2020;9:16.","DOI":"10.1167\/tvst.9.2.16"},{"key":"9880_CR21","first-page":"1","volume":"1","author":"MD Abr\u00e0moff","year":"2018","unstructured":"Abr\u00e0moff MD, Lavin PT, Birch M, Shah N, Folk JC. Pivotal trial of an autonomous AI-based diagnostic system for detection of diabetic retinopathy in primary care offices. NPJ Digit Med Nature Publishing Group. 2018;1:1\u20138.","journal-title":"NPJ Digit Med Nature Publishing Group"},{"key":"9880_CR22","unstructured":"Varma R. How AI benefits patients and physicians. Ophthalmol Times [Internet]. 2018; Available from: https:\/\/www.ophthalmologytimes.com\/view\/how-ai-benefits-patients-and-physicians"},{"key":"9880_CR23","volume-title":"Artificial intelligence for anterior segment diseases: emerging applications in ophthalmology","author":"DSJ Ting","year":"2020","unstructured":"Ting DSJ, Foo VHX, Yang LWY, Sia JT, Ang M, Lin H, et al. Artificial intelligence for anterior segment diseases: emerging applications in ophthalmology. Br J Ophthalmol: BMJ Publishing Group Ltd; 2020."},{"key":"9880_CR24","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1155\/2019\/8162567","volume":"2019","author":"A Lavric","year":"2019","unstructured":"Lavric A, Valentin P. KeratoDetect: keratoconus detection algorithm using convolutional neural networks. Comput Intell Neurosci. 2019;2019:1\u20139.","journal-title":"Comput Intell Neurosci"},{"key":"9880_CR25","unstructured":"Krizhevsky A, Sutskever I, Hinton GE. Imagenet classification with deep convolutional neural networks. Adv Neural Inf Process Syst. 2012. p. 1097\u2013105."},{"key":"9880_CR26","unstructured":"Iandola FN, Han S, Moskewicz MW, Ashraf K, Dally WJ, Keutzer K. SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and< 0.5 MB model size. arXiv Prepr arXiv160207360. 2016;"},{"key":"9880_CR27","doi-asserted-by":"crossref","unstructured":"Zhang X, Zhou X, Lin M, Sun J. Shufflenet: an extremely efficient convolutional neural network for mobile devices. Proc IEEE Conf Comput Vis pattern Recognit. 2018. p. 6848\u201356.","DOI":"10.1109\/CVPR.2018.00716"},{"key":"9880_CR28","doi-asserted-by":"crossref","unstructured":"Sandler M, Howard A, Zhu M, Zhmoginov A, Chen L-C. Mobilenetv2: inverted residuals and linear bottlenecks. Proc IEEE Conf Comput Vis pattern Recognit. 2018. p. 4510\u201320.","DOI":"10.1109\/CVPR.2018.00474"},{"key":"9880_CR29","unstructured":"Beale MH, Hagan MT, Demuth HB. Deep Learning ToolboxTM user\u2019s guide. MathWorks Inc. 2019;1\u201320."},{"key":"9880_CR30","unstructured":"Abd\u00fclh\u00fcssein NS. Building smart algorithm to extract features of topographic images of a human eye. MSc thesis, \u00dcniversitesi Fen Bilimleri Enstit\u00fcs\u00fc, Aksaray; 2018."},{"key":"9880_CR31","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1136\/bmjopen-2019-030833","volume":"9","author":"K Kamiya","year":"2019","unstructured":"Kamiya K, Ayatsuka Y, Kato Y, Fujimura F, Takahashi M, Shoji N, et al. Keratoconus detection using deep learning of colour-coded maps with anterior segment optical coherence tomography: a diagnostic accuracy study. BMJ Open. 2019;9:1\u20137.","journal-title":"BMJ Open"},{"key":"9880_CR32","doi-asserted-by":"publisher","first-page":"21","DOI":"10.1109\/MCAS.2006.1688199","volume":"6","author":"R Polikar","year":"2006","unstructured":"Polikar R. Ensemble based systems in decision making. IEEE Circuits Syst Mag. 2006;6:21\u201344.","journal-title":"IEEE Circuits Syst Mag"},{"key":"9880_CR33","doi-asserted-by":"crossref","unstructured":"Abr\u00e0moff MD, Lou Y, Erginay A, Clarida W, Amelon R, Folk JC, et al. Improved automated detection of diabetic retinopathy on a publicly available dataset through integration of deep learning. Invest Ophthalmol Vis Sci. The Association for Research in Vision and Ophthalmology; 2016;57:5200\u20136.","DOI":"10.1167\/iovs.16-19964"},{"key":"9880_CR34","doi-asserted-by":"crossref","unstructured":"Qian X, Patton EW, Swaney J, Xing Q, Zeng T. Machine learning on cataracts classification using SqueezeNet. 4th Int Conf Univers Village. IEEE; 2018. p. 1\u20133.","DOI":"10.1109\/UV.2018.8642133"},{"key":"9880_CR35","doi-asserted-by":"publisher","first-page":"2191","DOI":"10.1001\/jama.2013.281053","volume":"310","author":"WM Association","year":"2013","unstructured":"Association WM. World Medical Association Declaration of Helsinki: ethical principles for medical research involving human subjects. Jama American Medical Association. 2013;310:2191\u20134.","journal-title":"Jama American Medical Association"},{"key":"9880_CR36","unstructured":"Ueki R, Maeda N, Fuchihata M, Koh S, Kitaoka T, Nishida K. Differentiation of forme fruste keratoconus from normal cornea using parameters of corneal tomography, aberration, and biomechanics. Invest Ophthalmol Vis Sci. The Association for Research in Vision and Ophthalmology; 2014;55:3705."},{"key":"9880_CR37","doi-asserted-by":"publisher","unstructured":"Johnson JM, Khoshgoftaar TM. Survey on deep learning with class imbalance. J Big Data [Internet]. Springer International Publishing; 2019;6. Available from:\u00a0https:\/\/doi.org\/10.1186\/s40537-019-0192-5","DOI":"10.1186\/s40537-019-0192-5"},{"key":"9880_CR38","doi-asserted-by":"publisher","first-page":"321","DOI":"10.1613\/jair.953","volume":"16","author":"NV Chawla","year":"2002","unstructured":"Chawla NV, Bowyer KW, Hall LO, Kegelmeyer WP. SMOTE: synthetic minority over-sampling technique. J Artif Intell Res. 2002;16:321\u201357.","journal-title":"J Artif Intell Res"},{"key":"9880_CR39","volume-title":"Data mining: practical machine learning tools and techniques","author":"I Witten","year":"2005","unstructured":"Witten I, Frank E. Data mining: practical machine learning tools and techniques. 2nd ed. San Francisco: Morgan Kaufmann; 2005.","edition":"2"},{"key":"9880_CR40","doi-asserted-by":"crossref","unstructured":"Selvaraju RR, Cogswell M, Das A, Vedantam R, Parikh D, Batra D. Grad-cam: visual explanations from deep networks via gradient-based localization. Proc IEEE Int Conf Comput Vis. 2017. p. 618\u201326.","DOI":"10.1109\/ICCV.2017.74"},{"key":"9880_CR41","doi-asserted-by":"crossref","unstructured":"Panwar H, Gupta PK, Siddiqui MK, Morales-Menendez R, Bhardwaj P, Singh V. A deep learning and grad-CAM based color visualization approach for fast detection of COVID-19 cases using chest X-ray and CT-Scan images. Chaos, Solitons & Fractals. Elsevier; 2020;140:110190.","DOI":"10.1016\/j.chaos.2020.110190"},{"key":"9880_CR42","doi-asserted-by":"publisher","DOI":"10.1016\/j.ajo.2020.06.005","volume-title":"Corneal topography raw data classification using a convolutional neural network","author":"P Z\u00e9boulon","year":"2020","unstructured":"Z\u00e9boulon P, Debellemani\u00e8re G, Bouvet M, Gatinel D. Corneal topography raw data classification using a convolutional neural network. Am J Ophthalmol: Elsevier; 2020."},{"key":"9880_CR43","doi-asserted-by":"publisher","first-page":"358","DOI":"10.1111\/opo.12369","volume":"37","author":"JJ Rozema","year":"2017","unstructured":"Rozema JJ, Rodriguez P, Ruiz Hidalgo I, Navarro R, Tassignon M, Koppen C. SyntEyes KTC: higher order statistical eye model for developing keratoconus. Ophthalmic Physiol Opt Wiley Online Library. 2017;37:358\u201365.","journal-title":"Ophthalmic Physiol Opt Wiley Online Library"},{"key":"9880_CR44","doi-asserted-by":"crossref","unstructured":"Kuo B-I, Chang W-Y, Liao T-S, Liu F-Y, Liu H-Y, Chu H-S, et al. Keratoconus screening based on deep learning approach of corneal topography. Transl Vis Sci Technol. The Association for Research in Vision and Ophthalmology; 2020;9:53.","DOI":"10.1167\/tvst.9.2.53"},{"key":"9880_CR45","unstructured":"Sabour S, Frosst N, Hinton GE. Dynamic routing between capsules. Adv Neural Inf Process Syst. 2017. p. 3856\u201366."}],"container-title":["Cognitive Computation"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/s12559-021-09880-3.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/article\/10.1007\/s12559-021-09880-3\/fulltext.html","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/link.springer.com\/content\/pdf\/10.1007\/s12559-021-09880-3.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2022,9,9]],"date-time":"2022-09-09T07:56:38Z","timestamp":1662710198000},"score":1,"resource":{"primary":{"URL":"https:\/\/link.springer.com\/10.1007\/s12559-021-09880-3"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2021,6,16]]},"references-count":45,"journal-issue":{"issue":"5","published-print":{"date-parts":[[2022,9]]}},"alternative-id":["9880"],"URL":"https:\/\/doi.org\/10.1007\/s12559-021-09880-3","relation":{},"ISSN":["1866-9956","1866-9964"],"issn-type":[{"value":"1866-9956","type":"print"},{"value":"1866-9964","type":"electronic"}],"subject":[],"published":{"date-parts":[[2021,6,16]]},"assertion":[{"value":"15 August 2020","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"22 March 2021","order":2,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"16 June 2021","order":3,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}},{"order":1,"name":"Ethics","group":{"name":"EthicsHeading","label":"Declarations"}},{"value":"All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and\/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.","order":2,"name":"Ethics","group":{"name":"EthicsHeading","label":"Ethical Approval"}},{"value":"Informed consent was obtained originally\u00a0by Al-Amal ophthalmic centre at the time of data collection from all participants included in the study.","order":3,"name":"Ethics","group":{"name":"EthicsHeading","label":"Informed Consent"}},{"value":"The authors declare no competing interests.","order":4,"name":"Ethics","group":{"name":"EthicsHeading","label":"Conflict of Interest"}}]}}