<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE article PUBLIC "-//NLM//DTD JATS (Z39.96) Journal Publishing DTD v1.3 20210610//EN" "JATS-journalpublishing1-3.dtd">
<article article-type="research-article" dtd-version="1.3" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xml:lang="ru"><front><journal-meta><journal-id journal-id-type="publisher-id">helmholtzeyeinstitute</journal-id><journal-title-group><journal-title xml:lang="ru">Российский офтальмологический журнал</journal-title><trans-title-group xml:lang="en"><trans-title>Russian Ophthalmological Journal</trans-title></trans-title-group></journal-title-group><issn pub-type="ppub">2072-0076</issn><issn pub-type="epub">2587-5760</issn><publisher><publisher-name>Real time Publishers</publisher-name></publisher></journal-meta><article-meta><article-id pub-id-type="doi">10.21516/2072-0076-2025-18-3-supplement-8-11</article-id><article-id custom-type="elpub" pub-id-type="custom">helmholtzeyeinstitute-1915</article-id><article-categories><subj-group subj-group-type="heading"><subject>Research Article</subject></subj-group><subj-group subj-group-type="section-heading" xml:lang="ru"><subject>Статьи</subject></subj-group><subj-group subj-group-type="section-heading" xml:lang="en"><subject>Articles</subject></subj-group></article-categories><title-group><article-title>Диагностика офтальмологических и системных заболеваний по изображениям глазного дна с использованием мультимодальной трансформерной модели</article-title><trans-title-group xml:lang="en"><trans-title>Diagnostics of ophthalmological and systemic diseases from fundus images using a multimodal transformer model</trans-title></trans-title-group></title-group><contrib-group><contrib contrib-type="author" corresp="yes"><contrib-id contrib-id-type="orcid">https://orcid.org/0000-0001-5391-5229</contrib-id><name-alternatives><name name-style="eastern" xml:lang="ru"><surname>Аксенов</surname><given-names>К. Д.</given-names></name><name name-style="western" xml:lang="en"><surname>Aksenov</surname><given-names>K. D.</given-names></name></name-alternatives><bio xml:lang="ru"><p>Кирилл Дмитриевич Аксенов — генеральный директор; научный сотрудник</p><p>наб. им. Адмирала Серебрякова, д. 49, Новороссийск, Краснодарский край, 353905</p><p>ул. Карла Маркса, д. 20, Новороссийск, Краснодарский край, 353900</p></bio><bio xml:lang="en"><p>Kirill D. Aksenov — CEO; researcher</p><p>Admiral Serebryakov Emb., 49, Novorossiysk, Krasnodar Region, 353905</p><p>20, Karl Marx St., Novorossiysk, Krasnodar Region, 353900</p></bio><xref ref-type="aff" rid="aff-1"/></contrib><contrib contrib-type="author" corresp="yes"><contrib-id contrib-id-type="orcid">https://orcid.org/0000-0003-0885-1355</contrib-id><name-alternatives><name name-style="eastern" xml:lang="ru"><surname>Аксенова</surname><given-names>Л. Е.</given-names></name><name name-style="western" xml:lang="en"><surname>Aksenova</surname><given-names>L. E.</given-names></name></name-alternatives><bio xml:lang="ru"><p>Любовь Евгеньевна Аксенова — научный сотрудник</p><p>наб. им. Адмирала Серебрякова, д. 49, Новороссийск, Краснодарский край, 353905</p><p>ул. Карла Маркса, д. 20, Новороссийск, Краснодарский край, 353900</p></bio><bio xml:lang="en"><p>Lyubov E. Aksenova — researcher </p><p>Admiral Serebryakov Emb., 49, Novorossiysk, Krasnodar Region, 353905</p><p>20, Karl Marx St., Novorossiysk, Krasnodar Region, 353900</p></bio><email xlink:type="simple">axenovalubov@gmail.com</email><xref ref-type="aff" rid="aff-1"/></contrib></contrib-group><aff-alternatives id="aff-1"><aff xml:lang="ru"><institution>ООО «Пространство интеллектуальных решений»; Новороссийский политехнический институт (филиал) ФГБОУ ВПО «КубГТУ»</institution><country>Россия</country></aff><aff xml:lang="en"><institution>PREDICT SPACE LLC; Novorossiysk Polytechnic Institute (branch) of the Federal State Budgetary Educational Institution of Higher Professional Education “KubSTU”</institution><country>Russian Federation</country></aff></aff-alternatives><pub-date pub-type="collection"><year>2025</year></pub-date><pub-date pub-type="epub"><day>23</day><month>10</month><year>2025</year></pub-date><volume>18</volume><issue>3</issue><issue-title>ПРИЛОЖЕНИЕ</issue-title><fpage>8</fpage><lpage>11</lpage><permissions><copyright-statement>Copyright &amp;#x00A9; Аксенов К.Д., Аксенова Л.Е., 2025</copyright-statement><copyright-year>2025</copyright-year><copyright-holder xml:lang="ru">Аксенов К.Д., Аксенова Л.Е.</copyright-holder><copyright-holder xml:lang="en">Aksenov K.D., Aksenova L.E.</copyright-holder><license xml:lang="ru" license-type="creative-commons-attribution" xlink:href="https://creativecommons.org/licenses/by/4.0/" xlink:type="simple"><license-p>Данная работа распространяется под лицензией Creative Commons Attribution 4.0.</license-p></license><license xml:lang="en" license-type="creative-commons-attribution" xlink:href="https://creativecommons.org/licenses/by/4.0/" xlink:type="simple"><license-p>This work is licensed under a Creative Commons Attribution 4.0 License.</license-p></license></permissions><self-uri xlink:href="https://roj.igb.ru/jour/article/view/1915">https://roj.igb.ru/jour/article/view/1915</self-uri><abstract><p>Цель работы — оценка возможности диагностики офтальмологических и системных заболеваний по изображениям глазного дна с использованием мультимодальной трансформерной модели, обученной на открытом наборе данных.</p><sec><title>Материал и методы</title><p>Материал и методы. Для обучения и валидации использован открытый набор данных RFMiD, содержащий 3200 изображений глазного дна, размеченных относительно 29 классов заболеваний. В качестве модели использована предварительно обученная мультимодальная трансформерная архитектура, дообученная на этом наборе данных.</p></sec><sec><title>Результаты</title><p>Результаты. Модель показала стабильную сходимость и высокую точность при определении 29 классов заболеваний по изображениям глазного дна, достигнув AUC 0.9155 без признаков переобучения.</p></sec><sec><title>Заключение</title><p>Заключение. Полученные результаты демонстрируют высокую производительность мультимодальной модели на основе трансформенной архитектуры для задач многоклассовой классификации заболеваний по изображениям глазного дна.</p></sec></abstract><trans-abstract xml:lang="en"><sec><title>Purpose</title><p>Purpose: to evaluate of the potential for diagnosing ophthalmological and systemic diseases from fundus images using a multimodal transformer model trained on an open dataset.</p></sec><sec><title>Material and methods</title><p>Material and methods. An open RFMiD dataset containing 3200 fundus images annotated across 29 disease classes was used for training and validation. A pre-trained multimodal transformer architecture was used and fine-tuned on this dataset.</p></sec><sec><title>Results</title><p>Results. The model demonstrated stable convergence and high accuracy in identifying 29 disease classes from fundus images, achieving a test AUC of 0.9155 without signs of overfitting.</p></sec><sec><title>Conclusion</title><p>Conclusion. The obtained results show high performance of the multimodal transformer-based model for the task of multiclass disease classification from fundus images.</p></sec></trans-abstract><kwd-group xml:lang="ru"><kwd>глазное дно</kwd><kwd>набор данных</kwd><kwd>искусственный интеллект</kwd><kwd>нейронная сеть</kwd><kwd>классификация</kwd></kwd-group><kwd-group xml:lang="en"><kwd>fundus</kwd><kwd>dataset</kwd><kwd>artificial intelligence</kwd><kwd>neural network</kwd><kwd>classification</kwd></kwd-group><funding-group><funding-statement xml:lang="ru">исследование выполнено при финансовой поддержке Кубанского научного фонда, ООО «ПИР» в рамках проекта № НТИП-24.1/1 «Портативный оптический прибор для визуализации глазного дна со встроенными технологиями искусственного интеллекта».</funding-statement><funding-statement xml:lang="en">the research is carried out with the financial support of the Kuban Science Foundation; LLC PREDICT SPACE in the framework of the project Num. NTIP-24.1/1 “A portable optical device for visualizing the fundus with integrated artificial intelligence technologies”.</funding-statement></funding-group></article-meta></front><back><ref-list><title>References</title><ref id="cit1"><label>1</label><citation-alternatives><mixed-citation xml:lang="ru">Abràmoff M, Garvin M, Sonka M. Retinal imaging and image analysis. IEEE RevBiomed Eng. 2010; 3: 169–208. doi: 10.1109/RBME.2010.2084567</mixed-citation><mixed-citation xml:lang="en">Abràmoff M, Garvin M, Sonka M. Retinal imaging and image analysis. IEEE RevBiomed Eng. 2010; 3: 169–208. doi: 10.1109/RBME.2010.2084567</mixed-citation></citation-alternatives></ref><ref id="cit2"><label>2</label><citation-alternatives><mixed-citation xml:lang="ru">Issa P, Troeger E, Finger R, et al. Structure-function correlation of the human central retina. PLoS One. 2010 Sep 22; 5 (9): e12864. doi: 10.1371/journal.pone.0012864</mixed-citation><mixed-citation xml:lang="en">Issa P, Troeger E, Finger R, et al. Structure-function correlation of the human central retina. PLoS One. 2010 Sep 22; 5 (9): e12864. doi: 10.1371/journal.pone.0012864</mixed-citation></citation-alternatives></ref><ref id="cit3"><label>3</label><citation-alternatives><mixed-citation xml:lang="ru">Patton N, Aslam T, MacGillivray T, et al. Retinal image analysis: concepts, applications and potential. Prog Retin Eye Res. 2006 Jan; 25 (1): 99–127. doi: 10.1016/j.preteyeres.2005.07.001</mixed-citation><mixed-citation xml:lang="en">Patton N, Aslam T, MacGillivray T, et al. Retinal image analysis: concepts, applications and potential. Prog Retin Eye Res. 2006 Jan; 25 (1): 99–127. doi: 10.1016/j.preteyeres.2005.07.001</mixed-citation></citation-alternatives></ref><ref id="cit4"><label>4</label><citation-alternatives><mixed-citation xml:lang="ru">Li LY, Isaksen AA, Lebiecka-Johansen B, et al. Prediction of cardiovascular markers and diseases using retinal fundus images and deep learning: a systematic scoping review. Eur Heart J Digit Health. 2024 Sep 10; 5 (6): 660–9. doi: 10.1093/ehjdh/ztae068</mixed-citation><mixed-citation xml:lang="en">Li LY, Isaksen AA, Lebiecka-Johansen B, et al. Prediction of cardiovascular markers and diseases using retinal fundus images and deep learning: a systematic scoping review. Eur Heart J Digit Health. 2024 Sep 10; 5 (6): 660–9. doi: 10.1093/ehjdh/ztae068</mixed-citation></citation-alternatives></ref><ref id="cit5"><label>5</label><citation-alternatives><mixed-citation xml:lang="ru">Ting DSW, Cheung CY, Lim G, et al. Development and validation of a deep learning system for diabetic retinopathy and related eye diseases using retinal images from multiethnic populations with diabetes. JAMA. 2017 Dec 12; 318 (22): 2211–23. doi: 10.1001/jama.2017.18152</mixed-citation><mixed-citation xml:lang="en">Ting DSW, Cheung CY, Lim G, et al. Development and validation of a deep learning system for diabetic retinopathy and related eye diseases using retinal images from multiethnic populations with diabetes. JAMA. 2017 Dec 12; 318 (22): 2211–23. doi: 10.1001/jama.2017.18152</mixed-citation></citation-alternatives></ref><ref id="cit6"><label>6</label><citation-alternatives><mixed-citation xml:lang="ru">Pachade S, Porwal P, Thulkar D, et al. Retinal fundus multi-disease image dataset (RFMiD). IEEE Dataport. 2020 November, 25. doi:10.21227/s3g7-st65</mixed-citation><mixed-citation xml:lang="en">Pachade S, Porwal P, Thulkar D, et al. Retinal fundus multi-disease image dataset (RFMiD). IEEE Dataport. 2020 November, 25. doi:10.21227/s3g7-st65</mixed-citation></citation-alternatives></ref><ref id="cit7"><label>7</label><citation-alternatives><mixed-citation xml:lang="ru">Zhou Y, Chia MA, Wagner SK, et al. A foundation model for generalizable disease detection from retinal images. Nature. 2023; 622: 156–63.</mixed-citation><mixed-citation xml:lang="en">Zhou Y, Chia MA, Wagner SK, et al. A foundation model for generalizable disease detection from retinal images. Nature. 2023; 622: 156–63.</mixed-citation></citation-alternatives></ref></ref-list><fn-group><fn fn-type="conflict"><p>The authors declare that there are no conflicts of interest present.</p></fn></fn-group></back></article>
