Age-related macular degeneration (AMD) is a leading cause of vision loss among people aged 55 and older. AMD can be classified into early, intermediate, and late stages; and it is critical to predicting progression to late AMD. Here, we propose a framework that combines deep learning and survival analysis to automatically predict progression to late AMD from color fundus photographs. Experiments on AREDS dataset demonstrated that our framework achieved state-of-the-art performance.

Learning Objective: 1. Understand the problem of late Age-related macular degeneration (AMD) progression prediction.
2. Apply the deep learning and survival model on color fundus photographs.
3. Test the survival model using c-index.


Yifan Peng (Presenter)
National Institutes of Health

Tiarnan Keenan, National Eye Institute
Qingyu Chen, National Institutes of Health
Elvira Agrón, National Eye Institute
Wai Wong, National Eye Institute
Emily Chew, National Eye Institute
Zhiyong Lu, National Institutes of Health

Presentation Materials: