Home

Awesome

STSC

LICENSE Python PyTorch

Semantic-aware Texture-Structure Feature Collaboration for Underwater Image Enhancement [IEEE ICRA 2022]

By Di Wang, Long Ma, Risheng Liu, and Xin Fan

<div align=center> <img src="https://github.com/wdhudiekou/STSC/blob/master/Fig/network.png" width="80%"> </div>

Updates

[2022-05-23] Our paper is available online! [Paper]

Requirements

Get start

Please download the pretrained VGG model MyVGG.pt and vggfeature.pth and put them into the folder 'pretrain' Baidu Disk pretrained VGG model MyVGG.pt [code: c915] and a vggfeature.pth [code: 84zk]

  1. Check the dataset path in train.py, and then run:
       python train.py
    
    
  2. Check the model and image pathes in test_UIEB.py and then run:
       python test_UIEB.py
    

Please download the UIEB pretrained model and put it into folder './checkpoints/'

Baidu Disk UIEB pretrained model [code: mgey]

Dataset

Please download the following datasets:

Experimental Results

<div align=center> <img src="https://github.com/wdhudiekou/STSC/blob/master/Fig/UIEB.png" width="95%"> </div> <div align=center> <img src="https://github.com/wdhudiekou/STSC/blob/master/Fig/EUVP.png" width="95%"> </div> <div align=center> <img src="https://github.com/wdhudiekou/STSC/blob/master/Fig/RUIE.png" width="95%"> </div> <div align=center> <img src="https://github.com/wdhudiekou/STSC/blob/master/Fig/USOD.png" width="95%"> </div>

Citation

@InProceedings{Wang_22_ICRA,
  author    = {Di Wang and Long Ma and Risheng Liu and Xin Fan},
  title     = {Semantic-aware Texture-Structure Feature Collaboration for Underwater
               Image Enhancement},
  booktitle = {{IEEE} International Conference on Robotics and Automation {ICRA}},
  pages     = {4592--4598},
  year      = {2022}
}