README.md

Train on the StabStitch-D dataset

Modify the train_path and test_path in train_ssd.py and run:

python train_ssd.py

Train on TraditionalDataset

Modify the train_path and test_path in train_tra.py and run:

python train_tra.py

Generate spatial warps and save them for the warp smoothing model

Modify the test_path in test_ssd.py or test_tra.py and run:

python test_ssd.py

or

python test_tra.py

NOTE

1. We only generate the spatial warps for the training set.
2. The traditional dataset is not split into training and testing sets. We merely validate the robustness of StabStitch++ on it.