{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,17]],"date-time":"2026-09-17T07:47:53Z","timestamp":1789631273855,"version":"build-2803163510"},"reference-count":36,"publisher":"MDPI AG","issue":"2","license":[{"start":{"date-parts":[[2026,1,6]],"date-time":"2026-01-06T00:00:00Z","timestamp":1767657600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/501100003725","name":"National Research Foundation of Korea","doi-asserted-by":"crossref","award":["RS-2023-00249104"],"award-info":[{"award-number":["RS-2023-00249104"]}],"id":[{"id":"10.13039\/501100003725","id-type":"DOI","asserted-by":"crossref"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Diagnostics"],"abstract":"<jats:p>Background\/Objectives: Retinal optical coherence tomography (OCT) is essential for diagnosing ocular diseases, yet developing high-performing multiclass classifiers remains challenging due to limited labeled data and the computational cost of self-supervised pretraining. This study aims to address these limitations by introducing a curriculum-based self-supervised framework to improve representation learning and reduce computational burden for OCT classification. Methods: Two ensemble strategies were developed using progressive masked autoencoder (MAE) pretraining. We refer to this curriculum-based MAE framework as CurriMAE (curriculum-based masked autoencoder). CurriMAE-Soup merges multiple curriculum-aware pretrained checkpoints using weight averaging, producing a single model for fine-tuning and inference. CurriMAE-Greedy selects top-performing fine-tuned models from different pretraining stages and ensembles their predictions. Both approaches rely on one curriculum-guided MAE pretraining run, avoiding repeated training with fixed masking ratios. Experiments were conducted on two publicly available retinal OCT datasets, the Kermany dataset for self-supervised pretraining and the OCTDL dataset for downstream evaluation. The OCTDL dataset comprises seven clinically relevant retinal classes, including normal retina, age-related macular degeneration (AMD), diabetic macular edema (DME), epiretinal membrane (ERM), retinal vein occlusion (RVO), retinal artery occlusion (RAO), and vitreomacular interface disease (VID) and the proposed methods were compared against standard MAE variants and supervised baselines including ResNet-34 and ViT-S. Results: Both CurriMAE methods outperformed standard MAE models and supervised baselines. CurriMAE-Greedy achieved the highest performance with an area under the receiver operating characteristic curve (AUC) of 0.995 and accuracy of 93.32%, while CurriMAE-Soup provided competitive accuracy with substantially lower inference complexity. Compared with MAE models trained at fixed masking ratios, the proposed methods improved accuracy while requiring fewer pretraining runs and reduced model storage for inference. Conclusions: The proposed curriculum-based self-supervised ensemble framework offers an effective and resource-efficient solution for multiclass retinal OCT classification. By integrating progressive masking with snapshot-based model fusion, CurriMAE methods provide high performance with reduced computational cost, supporting their potential for real-world ophthalmic imaging applications where labeled data and computational resources are limited.<\/jats:p>","DOI":"10.3390\/diagnostics16020179","type":"journal-article","created":{"date-parts":[[2026,1,6]],"date-time":"2026-01-06T11:13:18Z","timestamp":1767697998000},"page":"179","update-policy":"https:\/\/doi.org\/10.3390\/mdpi_crossmark_policy","source":"Crossref","is-referenced-by-count":3,"title":["Efficient Ensemble Learning with Curriculum-Based Masked Autoencoders for Retinal OCT Classification"],"prefix":"10.3390","volume":"16","author":[{"ORCID":"https:\/\/orcid.org\/0009-0008-9454-8590","authenticated-orcid":false,"given":"Taeyoung","family":"Yoon","sequence":"first","affiliation":[{"name":"School of Bio-Health Convergence, College of Natural Sciences, Sungshin Women\u2019s University, Seoul 01133, Republic of Korea"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Daesung","family":"Kang","sequence":"additional","affiliation":[{"name":"School of Bio-Health Convergence, College of Natural Sciences, Sungshin Women\u2019s University, Seoul 01133, Republic of Korea"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1968","published-online":{"date-parts":[[2026,1,6]]},"reference":[{"key":"ref_1","doi-asserted-by":"crossref","first-page":"115","DOI":"10.1038\/nature21056","article-title":"Dermatologist-level classification of skin cancer with deep neural networks","volume":"542","author":"Esteva","year":"2017","journal-title":"Nature"},{"key":"ref_2","doi-asserted-by":"crossref","first-page":"2402","DOI":"10.1001\/jama.2016.17216","article-title":"Development and Validation of a Deep Learning Algorithm for Detection of Diabetic Retinopathy in Retinal Fundus Photographs","volume":"316","author":"Gulshan","year":"2016","journal-title":"JAMA"},{"key":"ref_3","doi-asserted-by":"crossref","first-page":"60","DOI":"10.1016\/j.media.2017.07.005","article-title":"A survey on deep learning in medical image analysis","volume":"42","author":"Litjens","year":"2017","journal-title":"Med. Image Anal."},{"key":"ref_4","doi-asserted-by":"crossref","first-page":"1299","DOI":"10.1109\/TMI.2016.2535302","article-title":"Convolutional Neural Networks for Medical Image Analysis: Full Training or Fine Tuning?","volume":"35","author":"Tajbakhsh","year":"2016","journal-title":"IEEE Trans. Med. Imaging"},{"key":"ref_5","doi-asserted-by":"crossref","unstructured":"Khalid, U., Kaya, M., and Alhajj, R. (2025). Improving Coronary Artery Disease Diagnosis in Cardiac MRI with Self-Supervised Learning. Diagnostics, 15.","DOI":"10.3390\/diagnostics15202618"},{"key":"ref_6","unstructured":"Chaitanya, K., Erdil, E., Karani, N., and Konukoglu, E. (2020, January 6\u201312). Contrastive learning of global and local features for medical image segmentation with limited annotations. Proceedings of the 34th International Conference on Neural Information Processing Systems, Vancouver, BC, Canada."},{"key":"ref_7","first-page":"857","article-title":"Self-Supervised Learning: Generative or Contrastive","volume":"35","author":"Liu","year":"2023","journal-title":"IEEE Trans. Knowl. Data Eng."},{"key":"ref_8","doi-asserted-by":"crossref","unstructured":"Xie, Z.D., Zhang, Z., Cao, Y., Lin, Y.T., Bao, J.M., Yao, Z.L., Dai, Q., and Hu, H. (2022, January 19\u201324). SimMIM: A Simple Framework for Masked Image Modeling. Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition, CVPR, New Orleans, LA, USA.","DOI":"10.1109\/CVPR52688.2022.00943"},{"key":"ref_9","doi-asserted-by":"crossref","unstructured":"He, K.M., Chen, X.L., Xie, S.N., Li, Y.H., Doll\u00e1r, P., and Girshick, R. (2022, January 19\u201324). Masked Autoencoders Are Scalable Vision Learners. Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition, CVPR, New Orleans, LA, USA.","DOI":"10.1109\/CVPR52688.2022.01553"},{"key":"ref_10","doi-asserted-by":"crossref","first-page":"108762","DOI":"10.1016\/j.engappai.2024.108762","article-title":"Dynamic background reconstruction via masked autoencoders for infrared small target detection","volume":"135","author":"Peng","year":"2024","journal-title":"Eng. Appl. Artif. Intell."},{"key":"ref_11","doi-asserted-by":"crossref","unstructured":"Dong, X.Y., Bao, J.M., Zhang, T., Chen, D.D., Zhang, W.M., Yuan, L., Chen, D., Wen, F., and Yu, N. (2022, January 23\u201327). Bootstrapped Masked Autoencoders for Vision BERT Pretraining. Proceedings of the Computer Vision\u2013ECCV 2022, Tel Aviv, Israel.","DOI":"10.1007\/978-3-031-20056-4_15"},{"key":"ref_12","first-page":"e3019","article-title":"CurriMAE: Curriculum learning based masked autoencoders for multi-labeled pediatric thoracic disease classification","volume":"11","author":"Yoon","year":"2025","journal-title":"PeerJ Comput. Sci."},{"key":"ref_13","unstructured":"Huang, G., Li, Y., Pleiss, G., Liu, Z., Hopcroft, J.E., and Weinberger, K.Q. (2017). Snapshot ensembles: Train 1, get m for free. arXiv."},{"key":"ref_14","doi-asserted-by":"crossref","first-page":"1122","DOI":"10.1016\/j.cell.2018.02.010","article-title":"Identifying Medical Diagnoses and Treatable Diseases by Image-Based Deep Learning","volume":"172","author":"Kermany","year":"2018","journal-title":"Cell"},{"key":"ref_15","doi-asserted-by":"crossref","unstructured":"Ardelean, A.I., Ardelean, E.R., and Marginean, A. (2025). Can YOLO Detect Retinal Pathologies? A Step Towards Automated OCT Analysis. Diagnostics, 15.","DOI":"10.3390\/diagnostics15141823"},{"key":"ref_16","doi-asserted-by":"crossref","first-page":"30321","DOI":"10.1038\/s41598-024-81127-1","article-title":"Evaluating deep learning models for classifying OCT images with limited data and noisy labels","volume":"14","author":"Miladinovic","year":"2024","journal-title":"Sci. Rep."},{"key":"ref_17","unstructured":"Devlin, J., Chang, M.W., Lee, K., and Toutanova, K. (2019, January 2\u20137). BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding. Proceedings of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL HLT 2019), Minneapolis, MN, USA."},{"key":"ref_18","doi-asserted-by":"crossref","first-page":"115","DOI":"10.1007\/978-3-031-73119-8_12","article-title":"Masked Image Modelling for Retinal OCT Understanding","volume":"15188","author":"Pissas","year":"2024","journal-title":"Ophthalmic Med. Image Anal. Omia"},{"key":"ref_19","doi-asserted-by":"crossref","first-page":"106817","DOI":"10.1016\/j.neunet.2024.106817","article-title":"GO-MAE: Self-supervised pre-training via masked autoencoder for OCT image classification of gynecology","volume":"181","author":"Wang","year":"2025","journal-title":"Neural Netw."},{"key":"ref_20","doi-asserted-by":"crossref","first-page":"112055","DOI":"10.1016\/j.engappai.2025.112055","article-title":"Self-distilled masked autoencoders for medical images","volume":"160","author":"Yoon","year":"2025","journal-title":"Eng. Appl. Artif. Intell."},{"key":"ref_21","doi-asserted-by":"crossref","unstructured":"Zhou, Z.-H. (2025). Ensemble Methods: Foundations and Algorithms, CRC Press. [2nd ed.].","DOI":"10.1201\/9781003587774"},{"key":"ref_22","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1007\/3-540-45014-9_1","article-title":"Ensemble methods in machine learning","volume":"Volume 1857","author":"Dietterich","year":"2000","journal-title":"Multiple Classifier Systems"},{"key":"ref_23","doi-asserted-by":"crossref","first-page":"123","DOI":"10.1023\/A:1018054314350","article-title":"Bagging predictors","volume":"24","author":"Breiman","year":"1996","journal-title":"Mach. Learn."},{"key":"ref_24","doi-asserted-by":"crossref","first-page":"105151","DOI":"10.1016\/j.engappai.2022.105151","article-title":"Ensemble deep learning: A review","volume":"115","author":"Ganaie","year":"2022","journal-title":"Eng. Appl. Artif. Intell."},{"key":"ref_25","unstructured":"Lakshminarayanan, B., Pritzel, A., and Blundell, C. (2017). Simple and Scalable Predictive Uncertainty Estimation using Deep Ensembles. Advances in Neural Information Processing Systems, NeurIPS."},{"key":"ref_26","unstructured":"Wen, Y., Tran, D., and Ba, J. (2020). BatchEnsemble: An Alternative Approach to Efficient Ensemble and Lifelong Learning. arXiv."},{"key":"ref_27","doi-asserted-by":"crossref","first-page":"106989","DOI":"10.1016\/j.bspc.2024.106989","article-title":"DEEP GD: Deep learning based snapshot ensemble CNN with EfficientNet for glaucoma detection","volume":"100","author":"Geetha","year":"2025","journal-title":"Biomed. Signal Process. Control."},{"key":"ref_28","doi-asserted-by":"crossref","first-page":"31232","DOI":"10.1038\/s41598-025-15704-3","article-title":"Integrating snapshot ensemble learning into masked autoencoders for efficient self-supervised pretraining in medical imaging","volume":"15","author":"Yoon","year":"2025","journal-title":"Sci. Rep."},{"key":"ref_29","doi-asserted-by":"crossref","unstructured":"Jang, D.H., Yun, S., and Han, D. (October, January 29). Model Stock: All We Need Is Just a Few Fine-Tuned Models. Proceedings of the Computer Vision-ECCV 2024, Milan, Italy.","DOI":"10.1007\/978-3-031-72784-9_12"},{"key":"ref_30","unstructured":"Wortsman, M., Ilharco, G., Gadre, S.Y., Roelofs, R., Gontijo-Lopes, R., Morcos, A.S., Namkoong, H., Farhadi, A., Carmon, Y., and Kornblith, S. (2022, January 17\u201323). Model soups: Averaging weights of multiple fine-tuned models improves accuracy without increasing inference time. Proceedings of the International Conference on Machine Learning, Baltimore, MD, USA."},{"key":"ref_31","doi-asserted-by":"crossref","first-page":"307","DOI":"10.1016\/j.ejca.2022.07.002","article-title":"Model soups improve performance of dermoscopic skin cancer classifiers","volume":"173","author":"Maron","year":"2022","journal-title":"Eur. J. Cancer"},{"key":"ref_32","doi-asserted-by":"crossref","first-page":"365","DOI":"10.1038\/s41597-024-03182-7","article-title":"OCTDL: Optical Coherence Tomography Dataset for Image-Based Deep Learning Methods","volume":"11","author":"Kulyabin","year":"2024","journal-title":"Sci. Data"},{"key":"ref_33","unstructured":"Loshchilov, I., and Hutter, F. (2016). SGDR: Stochastic Gradient Descent with Restarts. arXiv."},{"key":"ref_34","doi-asserted-by":"crossref","unstructured":"Xiao, J., Bai, Y., Yuille, A., and Zhou, Z. (2023, January 3\u20137). Delving into Masked Autoencoders for Multi-Label Thorax Disease Classification. Proceedings of the IEEE\/CVF Winter Conference on Applications of Computer Vision (WACV), Waikoloa, HI, USA.","DOI":"10.1109\/WACV56688.2023.00358"},{"key":"ref_35","doi-asserted-by":"crossref","unstructured":"He, K.M., Zhang, X.Y., Ren, S.Q., and Sun, J. (2016, January 27\u201330). Deep Residual Learning for Image Recognition. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, NV, USA.","DOI":"10.1109\/CVPR.2016.90"},{"key":"ref_36","unstructured":"Dosovitskiy, A., Beyer, L., Kolesnikov, A., Weissenborn, D., Zhai, X., Unterthiner, T., Dehghani, M., Minderer, M., Heigold, G., and Gelly, S. (2021). An image is worth 16x16 words: Transformers for image recognition at scale. arXiv."}],"container-title":["Diagnostics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.mdpi.com\/2075-4418\/16\/2\/179\/pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,1,8]],"date-time":"2026-01-08T05:22:29Z","timestamp":1767849749000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.mdpi.com\/2075-4418\/16\/2\/179"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,1,6]]},"references-count":36,"journal-issue":{"issue":"2","published-online":{"date-parts":[[2026,1]]}},"alternative-id":["diagnostics16020179"],"URL":"https:\/\/doi.org\/10.3390\/diagnostics16020179","relation":{},"ISSN":["2075-4418"],"issn-type":[{"value":"2075-4418","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,1,6]]}}}