Prototype-guided Salient Attention for Few-Shot Segmentation

This is the implementation of our paper Prototype-guided Salient Attention for Few-Shot Segmentation that has been accepted to The Visual Computer.

Requirements

• Python 3.7
• PyTorch 1.5.1
• cuda 10.1
• tensorboard 1.14

Datasets

• PASCAL-5ⁱ: VOC2012 + SBD

• COCO-20ⁱ: COCO2014

  Please see OSLSM and FWB for more details on datasets.

Model

Download the pre-trained backbones from BAM and put them into the PSANet/initmodel directory.

Download our trained base learners from BAM and put them under PSANet/initmodel/PSPNet.

Test and Train

You only need to configure the relevant content in the relevant script file to run

Training

./train.sh  

Testing

./test.sh

Visualization

References

This repo is mainly built based on BAM. Thanks for their great work!

Citation

If you find this project useful, please consider citing:

@article{li2024psanet,
  title={Psanet: prototype-guided salient attention for few-shot segmentation},
  author={Li, Hao and Huang, Guoheng and Yuan, Xiaochen and Zheng, Zewen and Chen, Xuhang and Zhong, Guo and Pun, Chi-Man},
  journal={The Visual Computer},
  pages={1--15},
  year={2024},
  publisher={Springer}
}