TransRA

TransRA: transformer and residual attention fusion for single remote sensing image dehazing(https://doi.org/10.1007/s11045-022-00835-x)

Dependencies and Installation

• python3.8.5
• anaconda
• PyTorch =1.8.0
• NVIDIA GPU+CUDA
• numpy
• matplotlib
• tensorboardX(optional)

Pretrained Weights & Dataset

• Download our model weights.
• Download Haze_1k dataset
• Download RICE dataset

Dataset

Train

python train.py --data_dir data/Haze_1k/thick -train_batch_size 2 --model_save_dir train_result

Test

python test.py --model_save_dir results

Qualitative Results

Quantitative comparisons over SateHaze1k for different methods:

Citation

If you use any part of this code, please kindly cite

@article{Dong2022,
  title={TransRA: transformer and residual attention fusion for single remote sensing image dehazing},
  author={Dong, Pengwei, Wang, Bo},
  journal={Multidimensional Systems and Signal Processing},
  url={https://doi.org/10.1007/s11045-022-00835-x},
  year={2022}
}