"M-FunFound: A Minimally Supervised Fundus Foundation Model for Multiple Tasks"

Please contact balinda@snu.ac.kr or boajang97@gmail.com if you have any questions.

Main repository: https://github.com/Jang-Boa/Research-Foundation-Retina

_The repository is currently empty. All codes will be uploaded after publication. All the update of this repository will be shown up in Main repository first.

Key Features

• M-FunFound is pre-trained on 0.1 million retinal images with super-vised learning.
• M-FunFound can be efficiently adapted to customised tasks.

Fine-tuning with Fundus and ImageNet+Fundus weights

• "The current maintenance and GitHub update are scheduled. The model will be released after the paper is published."

📃Citation

If you find this repository useful, please consider citing this paper:

License

Code and Model Weights are only can used for Research.