Usage

This is the official code of our paper "Salient Object Detection Combining a Self-Attention Module and a Feature Pyramid Network".

1. Download the datasets

Download the following datasets and unzip them into data folder.

• DUTS dataset. The .lst file for training is data/DUTS/DUTS-TR/train_pair.lst.

• Datasets for testing.

2. Download the pre-trained models for backbone

Download the following pre-trained models GoogleDrive | BaiduYun (pwd: 27p5) into dataset/pretrained folder.

3. Train

1. Set the --train_root and --train_list path in train.sh correctly.

2. We demo using ResNet-50 as network backbone and train with a initial lr of 5e-5 for 24 epoches, which is divided by 10 after 15 epochs.

./train.sh

3. After training the result model will be stored under results/run-* folder.

4. Test

python main.py --mode='test' --model='results/run-*/models/final.pth' --test_fold='results/run-*-sal-e' --sal_mode='e'

All results saliency maps will be stored under results/run-* folders in .png formats.

5. Evaluation results

https://github.com/NathanUA/Binary-Segmentation-Evaluation-Tool. The Code was used for evaluation in CVPR 2019 paper 'BASNet: Boundary-Aware Salient Object Detection code', Xuebin Qin, Zichen Zhang, Chenyang Huang, Chao Gao, Masood Dehghan and Martin Jagersand.

BibTex

To cite this code for publications - please use:

@article{ren2020salient,
  title={Salient Object Detection Combining a Self-attention Module and a Feature Pyramid Network},
  author={Ren, Guangyu and Dai, Tianhong and Barmpoutis, Panagiotis and Stathaki, Tania},
  journal={Electronics},
  volume={9},
  number={10},
  pages={1702},
  year={2020},
  publisher={Multidisciplinary Digital Publishing Institute}
}