RGB-D-Saliency-Detection

Usage

This is the official code of our paper "Progressive multi-scale fusion network for RGB-D salient object detection".

1. Train

1. Set the --datasets_root_train and --datasets_root_test path in config.py.

2. Run train.py.

3. After training the result model will be stored under model* folder.

2. Test

python generate_salmap.py --exp_name='./'

3. Evaluation results

Run python test_metric_score.py

BibTex

To cite this code for publications - please use:

@article{ren2022progressive,
  title={Progressive multi-scale fusion network for RGB-D salient object detection},
  author={Ren, Guangyu and Xie, Yanchun and Dai, Tianhong and Stathaki, Tania},
  journal={Computer Vision and Image Understanding},
  volume={223},
  pages={103529},
  year={2022},
  publisher={Elsevier}
}