RMFormer

Reparameterized Multi-scale Transformer for Deformable Retinal Image Registration

Paper Link will be available soon.

Datasets

We used OASIS dataset processed by Junyu Chen [here].

Set train_dir and val_dir to your path in train_RMFormer.py line 32-33.

Then, simply run python train_RMFormer.py will start the training process.

Set test_dir to your path in infer_RMFormer.py line 45.

Then, simply run python infer_RMFormer.py will start the testing process.

This repo is heavily based on Junyu Chen's code (junyuchen245/TransMorph_Transformer_for_Medical_Image_Registration: TransMorph). Thanks for their contribution!