WFBNet

Code for Pacific Graphics 2020 paper "Coarse to Fine: Weak Feature Boosting Network for Salient Object Detection"

Prerequisites

• Python 3.5
• Pytorch 1.3
• OpenCV 4.0
• Numpy 1.15
• TensorboardX
• Apex

Download dataset

Download the following datasets and unzip them into data folder

• PASCAL-S
• ECSSD
• HKU-IS
• DUT-OMRON
• DUTS

Pretrained model for backbone

Please download the pretrained model for backbone into res folder

Training

    cd src
    python3 train.py

Testing

    cd src
    python3 test.py ${epoch_you_wanna_test}

Saliency maps & Trained model

• saliency maps: Baidu(iydw) Google
• trained model: Baidu(xrdj) Google
• If you want to test using our trained model, you can just download and unzip it to folder 'src', then run

  cd src     
  python3 test.py 33

Evaluation

• To evaluate the performace, please use MATLAB to run main.m

    cd eval
    matlab
    main

Citation

@article {WFBNet, 
  author = {Zhang, Chenhao and Gao, Shanshan and Pan, Xiao and Wang, Yuting and Zhou, Yuanfeng}, 
  title = {{Coarse to Fine:Weak Feature Boosting Network for Salient Object Detection}}, 
  journal = {Computer Graphics Forum}, 
  year = {2020}, 
  editor = {Eisemann, Elmar and Jacobson, Alec and Zhang, Fang-Lue}, 
  volume = {39}, 
  number = {7}, 
  publisher = {The Eurographics Association and John Wiley & Sons Ltd.}, 
  pages = {411-420}, 
  DOI = {10.1111/cgf.14155} 
}

Thanks to F3Net