data and checkpoints

BENCHMARK.md

@@ -0,0 +1,45 @@
+## RETFound - Benckmark data split and model checkpoints
+
+
+Here we provide data split and model checkpoints to facilitate model comparison. Please remember to accordingly change the data_path, task, and nb_classes for model fine-tuning and evaluation.
+
+### Data split
+| Dataset | Download Link 1 |   Download Link 2 |
+| ------------- | ------------------ |------------------ |
+| APTOS2019 | [Google Drive](https://drive.google.com/file/d/162YPf4OhMVxj9TrQH0GnJv0n7z7gJWpj/view?usp=sharing)      |      |
+| MESSIDOR2 | [Google Drive](https://drive.google.com/file/d/1vOLBUK9xdzNV8eVkRjVdNrRwhPfaOmda/view?usp=sharing)        |   |
+| IDRID | [Google Drive](https://drive.google.com/file/d/1c6zexA705z-ANEBNXJOBsk6uCvRnzmr3/view?usp=sharing)        |               |
+| PAPILA | [Google Drive](https://drive.google.com/file/d/1JltYs7WRWEU0yyki1CQw5-10HEbqCMBE/view?usp=sharing)      |     |
+| Glaucoma_fundus | [Google Drive](https://drive.google.com/file/d/18vSazOYDsUGdZ64gGkTg3E6jiNtcrUrI/view?usp=sharing)        |  |
+| JSIEC | [Google Drive](https://drive.google.com/file/d/1q0GFQb-dYwzIx8AwlaFZenUJItix4s8z/view?usp=sharing)        |              |
+| Retina | [Google Drive](https://drive.google.com/file/d/1vdmjMRDoUm9yk83HMArLiPcLDk_dm92Q/view?usp=sharing)      |    |
+| OCTID | [Google Drive](https://drive.google.com/file/d/1I7nAvbkJG4UF29J3HcyIW53rVEFcKRgm/view?usp=sharing)        |   |
+
+
+
+### Model checkpoints
+| Dataset | Download Link 1 |   Download Link 2 |
+| ------------- | ------------------ |------------------ |
+| APTOS2019 | [Google Drive](https://drive.google.com/drive/folders/16kL5V-1U7ACc-68PSHjAq6vyXRJvUoq3?usp=sharing)      |      |
+| MESSIDOR2 | [Google Drive](https://drive.google.com/drive/folders/1OTBRAHNbaytpwzwMHw9SWrltJouEEuxF?usp=sharing)        |   |
+| IDRID | [Google Drive](https://drive.google.com/drive/folders/18Ml-B7nhejK4rnNG8upjqIARSlMP5kUc?usp=sharing)        |              |
+| PAPILA | [Google Drive](https://drive.google.com/drive/folders/1cHOX6C4NQVi9B6n-7Bxxg7b4-wdI4c73?usp=sharing)      |      |
+| Glaucoma_fundus | [Google Drive](https://drive.google.com/drive/folders/10JbanmVxjyX6mghXbxGnGVX1p9nwqsja?usp=sharing)        |   |
+| JSIEC | [Google Drive](https://drive.google.com/drive/folders/1eosdBXsONUy49cwDO80AOzDHkHiPNJvv?usp=sharing)        |              |
+| Retina | [Google Drive](https://drive.google.com/drive/folders/1n7mXxN-ZUKauOrAlBAiF2E_36F6f0wZD?usp=sharing)      |     |
+| OCTID | [Google Drive](https://drive.google.com/drive/folders/14SQdLuIxfkiqz_zmpvNkd9Ka4NTW3Fml?usp=sharing)        |   |
+
+
+
+### Official websites
+
+APTOS2019: https://www.kaggle.com/competitions/aptos2019-blindness-detection/data
+MESSIDOR2: https://www.adcis.net/en/third-party/messidor2/
+IDRID: https://ieee-dataport.org/open-access/indian-diabetic-retinopathy-image-dataset-idrid
+PAPILA: https://figshare.com/articles/dataset/PAPILA/14798004/1
+Glaucoma_fundus: https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/1YRRAC
+JSIEC: https://zenodo.org/records/3477553
+Retina: https://www.kaggle.com/datasets/jr2ngb/cataractdataset
+OCTID: https://borealisdata.ca/dataverse/OCTID
+
+

README.md

@@ -17,6 +17,7 @@ Keras version implemented by Yuka Kihara can be found [here](https://github.com/
 
 ### 🎉News
 
+- 🐉2024/01: [Data split and model checkpoints]((BENCHMARK.md)) for public datasets are now online!
 - 🎄2023/12: [Colab notebook](https://colab.research.google.com/drive/1_X19zdMegmAlqPAEY0Ao659fzzzlx2IZ?usp=sharing) is now online - free GPU & simple operation!!!
 - 2023/09: a [visualisation demo](https://github.com/rmaphoh/RETFound_MAE/blob/main/RETFound_visualize.ipynb) is added
 - 2023/10: change the hyperparameter of [input_size](https://github.com/rmaphoh/RETFound_MAE#:~:text=finetune%20./RETFound_cfp_weights.pth%20%5C-,%2D%2Dinput_size%20224,-For%20evaluation%20only) for any image size
@@ -60,7 +61,7 @@ To fine tune RETFound on your own data, follow these steps:
 </tr>
 </tbody></table>
 
-2. Organise your data into this directory structure (using IDRiD as an [example](Example.ipynb))
+2. Organise your data into this directory structure (Public datasets used in this study can be [downloaded here](BENCHMARK.md))
 
 <p align="left">
   <img src="./pic/file_index.jpg" width="160">
@@ -87,7 +88,7 @@ python -m torch.distributed.launch --nproc_per_node=1 --master_port=48798 main_f
 ```
 
 
-4. For evaluation only
+4. For evaluation only (download data and model checkpoints [here](BENCHMARK.md); change the path below)
 
 
 ```
@@ -154,7 +155,9 @@ If you find this repository useful, please consider citing this paper:
   title={A foundation model for generalizable disease detection from retinal images},
   author={Zhou, Yukun and Chia, Mark A and Wagner, Siegfried K and Ayhan, Murat S and Williamson, Dominic J and Struyven, Robbert R and Liu, Timing and Xu, Moucheng and Lozano, Mateo G and Woodward-Court, Peter and others},
   journal={Nature},
-  pages={1--8},
+  volume={622},
+  number={7981},
+  pages={156--163},
   year={2023},
   publisher={Nature Publishing Group UK London}
 }