WestCityInstitute / AHDRNet

Attention-guided Network for Ghost-free High Dynamic Range Imaging

AHDRNet

Attention-guided Network for Ghost-free High Dynamic Range Imaging (AHDR)

Qingsen Yan*, Dong Gong*, Qinfeng Shi, Anton van den Hengel, Chunhua Shen, Ian Reid, Yanning Zhang. In IEEE Conference on Compute rVision and Pattern Recognition (CVPR), 2019:1751-1760. (* Equall contribution) [Paper][Project]

Requirements

• Python 2.7
• PyTorch 0.3.1 (tested with 0.3.1)
• MATLAB (for data preparation)

Usage

Data preparation

1. Download data from [dataset]
2. Move the dataset into ./GenerH5Data/TrainingData.
3. Run ./GenerH5Data/PrepareData.m

Testing

1. Install this repository and the required packages. A pretrained model is in ./trained-model.
2. Prepare dataset.
   1. Download dataset.
   2. Move the dataset into ./dataset.
   3. Processed dataset can be obtained by running the corresponding script in ./GenerH5Data/PrepareData.m.
3. Run python script_testing.py files.

Training

1. Prepare dataset.
   1. Download dataset.
   2. Move the dataset into ./dataset.
   3. Processed dataset can be obtained by running the corresponding script in ./GenerH5Data/PrepareData.m.
2. Run python script_training.py files.

Examples of the Results

Examples of the Estimated Attention Maps

Citation

If you use this code for your research, please cite our paper.

@article{yan2021dual,
  title={Dual-attention-guided network for ghost-free high dynamic range imaging},
  author={Yan, Qingsen and Gong, Dong and Shi, Javen Qinfeng and van den Hengel, Anton and Shen, Chunhua and Reid, Ian and Zhang, Yanning},
  journal={International Journal of Computer Vision},
  pages={1--19},
  year={2021},
  publisher={Springer}
}
@article{yan2019attention,
  title={Attention-guided Network for Ghost-free High Dynamic Range Imaging},
  author={Yan, Qingsen and Gong, Dong and Shi, Qinfeng and Hengel, Anton van den and Shen, Chunhua and Reid, Ian and Zhang, Yanning},
  journal={IEEE Conference on Compute rVision and Pattern Recognition (CVPR)},
  year={2019}
  pages={1751-1760}
}