A Feature Refinement and Adaptive Generative Adversarial Network for Thermal Infrared Image Colorization

• python 3.7
• torch 1.13.1
• torchvision 0.14.1
• dominate
• visdom

We provide the pretrained model: Google

python train.py --dataroot [dataset root] --name [experiment_name] --phase train --which_epoch latest

python test.py --dataroot [dataset root] --name [experiment_name] --phase test --which_epoch latest

We only uploaded the paired strategy, the remaining strategies can be found in CycleGAN and CUT.

This code heavily borrowes from MUGAN.