STAGE 2023 (2nd Place)

Code for Structural-Functional Transition in Glaucoma Assessment Challenge 2023 (STAGE 2023)

Preparing

1. Clone this repo:

git clone https://github.com/Chen-Ziyang/STAGE2023.git
cd STAGE2023

2. Download the pre-trained weights and move them into model_checkpoints folder of each task, which can be found in this Google Drive Link.

3. Download the dataset from the official website.

4. Create the experimental environment in your own way or download ours from this Google Drive Link.

Training

1. Task 1: Prediction of Mean Deviation

python STAGE2023_Task1/train_MixLoss.py

2. Task 2: Sensitivity map prediction

python STAGE2023_Task2/train_MixLoss.py

3. Task 3: Pattern deviation probability map prediction

python STAGE2023_Task3/train.py

Inference

1. Task 1: Prediction of Mean Deviation

python STAGE2023_Task1/test.py

2. Task 2: Sensitivity map prediction

python STAGE2023_Task2/test.py

3. Task 3: Pattern deviation probability map prediction

python STAGE2023_Task3/test.py