liuquande / FedDG-ELCFS

[CVPR'21] FedDG: Federated Domain Generalization on Medical Image Segmentation via Episodic Learning in Continuous Frequency Space
240 stars 34 forks source link

Questions about data augmentation in fundus dataset. #6

Closed BurningFr closed 3 years ago

BurningFr commented 3 years ago

Hi, In your released code, I cannot see the data augmentation operations in train_ELCFS.py which are introduced in your paper. We can only see you use the ToTensor() operation, which is weird.

liuquande commented 3 years ago

Hi,

Thanks for the interest.

We performed the data augmentation offline and saved the augmented data as 'npy' format, in order to save the time cost for data loading and speed up the federated learning process.