Awesome

CCFENet-PyTorch (TCSVT-2022)

IEEE TCSVT 2022: Cross-Collaborative Fusion-Encoder Network for Robust RGB-Thermal Salient Object Detection.

Requirements

• Pytorch 1.3.0+

• Torchvision

• PIL

• Numpy

VT1000-D Dataset

The VT1000-D Dataset used in the paper can be downloaded from here [code: lt62]

• Containing 5 different degrees of distortion.

• Containing 12 types of algorithmically generated degradations from blur (defocus, motion, and Gaussian), noise (shot, Gaussian, and impulse), digital (brightness, contrast, and jpeg compression), and weather (snow, frost, and fog) categories.

Results

• RGB-Thermal saliency maps mentioned in the paper can be downloaded from here [code: gprv]

• RGB-Depth saliency maps mentioned in the paper can be downloaded from here [code: qoc7]

• The results of challenging scenarios mentioned in the paper can be downloaded from here [code: fteh]

• The saliency results can be evaluated by using the tool in Matlab

Testing

• Download the trained model weight from here [code: ij0a]

• Modify your test_root in test.py

• Test the CCFENet: python test.py

Citation

Please consider citing our work if you use this repository in your research.

@ARTICLE{CCFENet_TCSVT22,
  author={Liao, Guibiao and Gao, Wei and Li, Ge and Wang, Junle and Kwong, Sam},
  journal={IEEE Transactions on Circuits and Systems for Video Technology}, 
  title={Cross-Collaborative Fusion-Encoder Network for Robust RGB-Thermal Salient Object Detection}, 
  year={2022},
  volume={32},
  number={11},
  pages={7646-7661},
  }

Acknowlogdement

This repository is built with the help of the following projects for academic use only:

• PyTorch

• JL-DCF