elliotvincent/b-FLAIR-test

b-FLAIR-test: Building Change Detection Evaluation Dataset Dataset Description b-FLAIR-test is an evaluation dataset for building change detection, containing 1,730 annotated image pairs with binary building change masks. This dataset is designed for in-domain evaluation of methods trained on b-FLAIR in particular and provides a rigorous benchmark for bi-temporal building change detection in general. Project page: https://xavibou.github.io/CDviaWTS/… See the full description on the dataset page: https://huggingface.co/datasets/elliotvincent/b-FLAIR-test.