This is the official repository for the MARIO (Monitoring AMD progression in OCT) Challenge. Here you'll find guidance on how to participate and submit your solutions to Codabench.
- Evaluation scripts
- Docker example for final phase submissions
- Access to the dataset on Zenodo
The first task focuses on pairs of 2D slices (B-scans) from two consecutive OCT acquisitions. The goal is to classify the evolution between these two slices (before and after), which clinicians typically examine side by side on their screens.
The second task operates at the 2D slice level. The goal is to predict the future evolution within 3 months for patients enrolled in anti-VEGF treatment plans. While Task 1 aims to automate the initial analysis step (decision support), Task 2 aims to automate the complete analysis process (autonomous AI).
🔔 Note: Only teams with valid submissions for both tasks will be considered for final ranking and rewards.
- Create a CodaBench account if you don't have one
- Download and complete the participation form
- Send the form to rachid.zeghlache@univ-brest.fr with "MARIO 2024 Challenge" in the subject
- Name your form as "MARIO 2024 Data Challenge Participation Form_team_name.pdf"
- Once verified, you'll receive a download link within 48 working hours
- Alternatively, access the public dataset on Zenodo
- Preliminary Phase: Submit your CSV predictions for the validation set
- Final Phase: Submit a container of your code to the same address with subject "Container solution [Team_name]"
Certainly! I've updated the message to include how to cite the research paper, using the provided information to format a standard BibTeX entry and general citation guidance.
We're pleased to announce the publication of our comprehensive analysis of the MARIO AMD Progression Challenge and the corresponding dataset is now publicly available. This research focuses on the application of deep learning for the assessment of retinal degeneration, specifically Age-related Macular Degeneration (AMD) progression.
For detailed insights into the methods, results, and analysis from the challenge, please refer to the pre-print on ArXiv:
- Title: Deep Learning for Retinal Degeneration Assessment: A Comprehensive Analysis of the MARIO AMD Progression Challenge
- Authors: Rachid Zeghlache et al.
- Year: 2025
-
ArXiv ID:
$2506.02976$ - Link: https://arxiv.org/abs/2506.02976
If you use the findings, methods, or resources presented in this paper, please cite the work using the following information:
@misc{zeghlache2025deeplearningretinaldegeneration,
title={Deep Learning for Retinal Degeneration Assessment: A Comprehensive Analysis of the MARIO AMD Progression Challenge},
author={Rachid Zeghlache and Ikram Brahim and Pierre-Henri Conze and Mathieu Lamard and Mohammed El Amine Lazouni and Zineb Aziza Elaouaber and Leila Ryma Lazouni and Christopher Nielsen and Ahmad O. Ahsan and Matthias Wilms and Nils D. Forkert and Lovre Antonio Budimir and Ivana Matovinović and Donik Vršnak and Sven Lončarić and Philippe Zhang and Weili Jiang and Yihao Li and Yiding Hao and Markus Frohmann and Patrick Binder and Marcel Huber and Taha Emre and Teresa Finisterra Araújo and Marzieh Oghbaie and Hrvoje Bogunović and Amerens A. Bekkers and Nina M. van Liebergen and Hugo J. Kuijf and Abdul Qayyum and Moona Mazher and Steven A. Niederer and Alberto J. Beltrán-Carrero and Juan J. Gómez-Valverde and Javier Torresano-Rodríquez and Álvaro Caballero-Sastre and María J. Ledesma Carbayo and Yosuke Yamagishi and Yi Ding and Robin Peretzke and Alexandra Ertl and Maximilian Fischer and Jessica Kächele and Sofiane Zehar and Karim Boukli Hacene and Thomas Monfort and Béatrice Cochener and Mostafa El Habib Daho and Anas-Alexis Benyoussef and Gwenolé Quellec},
year={2025},
eprint={2506.02976},
archivePrefix={arXiv},
primaryClass={cs.CV},
url={https://arxiv.org/abs/2506.02976},
}Zeghlache, R., Brahim, I., Conze, P.-H., et al. (2025). Deep Learning for Retinal Degeneration Assessment: A Comprehensive Analysis of the MARIO AMD Progression Challenge. arXiv. https://arxiv.org/abs/2506.02976
The dataset used in the MARIO AMD Progression Challenge is hosted on Zenodo and is available for further research and benchmarking:
-
Dataset Record:
$15270469$ - Link: https://zenodo.org/records/15270469
We encourage the research community to utilize these resources to advance the field of ophthalmic image analysis and deep learning for retinal disease assessment.