metadata
dataset_info:
features:
- name: imageA
dtype: image
- name: imageB
dtype: image
- name: label
dtype: image
splits:
- name: train
num_bytes: 112278146.48
num_examples: 4558
- name: test
num_bytes: 37388998.684
num_examples: 1519
- name: val
num_bytes: 37425501.773
num_examples: 1519
download_size: 186554180
dataset_size: 187092646.937
Dataset Card for "GVLM_Cropped_256"
This is an official release of the GVLM-CD dataset. In this version, we cropped the images into patches of size 256*256.
If you use GVLM-CD in a scientific publication, we would appreciate using the following citations:
@article{zhang2023cross,
title={Cross-domain landslide mapping from large-scale remote sensing images using prototype-guided domain-aware progressive representation learning},
author={Zhang, Xiaokang and Yu, Weikang and Pun, Man-On and Shi, Wenzhong},
journal={ISPRS Journal of Photogrammetry and Remote Sensing},
volume={197},
pages={1--17},
year={2023},
publisher={Elsevier}
}