Towards SAR-to-Optical Image Translation via Perception Correlative Learning and Global-Local Feature Collaborative
- python 3.10
- torch 2.3.1
- CUDA 11.8
- dominate
- visdom
python train.py --dataroot [] --name [] --model [] --gpu_ids 0 --lambda_spatial 10 --lambda_gradient 0 --attn_layers 4,7,9 --loss_mode cos --gan_mode lsgan --display_port 8097 --direction AtoB --patch_size 64
python test.py --dataroot [] --checkpoints_dir ./checkpoints --name [] --model [] --num_test []
This code heavily borrowes from CUT,F/LSeSim, and KAN-CUT.
@article{chen2026towards,
title={Towards SAR-to-optical image translation via perception correlative learning and global-local feature collaboration},
author={Chen, Yu and Zhan, Weida and Jiang, Yichun and Zhu, Depeng and Xu, Xiaoyu and Guo, Jinxin and Hao, Ziqiang},
journal={Pattern Recognition},
pages={114186},
year={2026},
publisher={Elsevier}
}