/DPNet

Code for the Neurocomputing 2020 paper "Dual pyramid network for salient object detection"

Primary LanguagePythonMIT LicenseMIT

Dual pyramid network for salient object detection

by Xuemiao Xu^, Jiaxing Chen^, Huaidong Zhang*, and Guoqiang Han* (^ joint 1st author, * joint corresponding author)[paper link]

This implementation is written by Jiaxing Chen at the South China University of Technology.

Citation

@article{xu2020dual,

     title={Dual pyramid network for salient object detection},

     author={Xu, Xuemiao and Chen, Jiaxing and Zhang, Huaidong and Han, Guoqiang},

     journal={Neurocomputing},

     volume={375},

     pages={113--123},

     year={2020},

     publisher={Elsevier}
}

Saliency Map

The results of DPNet on six RGB saliency datasets (ECSSD, HKU-IS, PASCAL-S, SOD, DUT-OMRON, DUTS-TE) and three RGB-D saliency datasets (NLPR, NJUD, STEREO) can be found at Google Drive.

Trained Model

You can download the trained model which is reported in our paper at Google Drive.

Requirement

  • Python 2.7
  • PyTorch 0.4.0
  • torchvision
  • numpy
  • Cython
  • pydensecrf (here to install)

Training

  1. Set the path of pretrained resnet model in resnet/config.py
  2. Set the path of DUTS-TR dataset in config.py
  3. Run by python train.py

Hyper-parameters of training were gathered at the beginning of train.py and you can conveniently change them as you need.

Testing

  1. Set the path of six benchmark datasets in config.py
  2. Put the trained model in ckpt/dpnet
  3. Run by python infer.py

Settings of testing were gathered at the beginning of infer.py and you can conveniently change them as you need.

Dataset links