RGB-D Saliency Datasets

Dataset Overview

Dataset Name

• RGBD‑SOD‑datasets

Dataset Content

• Contains multiple RGB‑D saliency detection datasets, provided as ready‑to‑download packages.

Dataset Composition

• Test Set: All test data, extraction code b2p2.
• Training Set 1: Includes NJUD and NLPR, extraction code 76gu.
• Training Set 2: Includes NJUD, NLPR, and DUT, extraction code 201p.

Detailed Information

Usage Tips

• All data are uniformly resized to 256 × 256 px for training.

Citation

• If you use the DUTLF‑Depth dataset, please cite:

  @inproceedings{piao2019depth,
    title={Depth‑induced multi‑scale recurrent attention network for saliency detection},
    author={Piao, Yongri and Ji, Wei and Li, Jingjing and Zhang, Miao and Lu, Huchuan},
    booktitle={Proceedings of the IEEE/CVF International Conference on Computer Vision},
    pages={7254--7263},
    year={2019}
  }
  

Evaluation Metrics

• Provides comprehensive saliency object detection metrics, including E‑measure, S‑measure, weighted F‑measure, MAE score and PR curve. Evaluation toolbox: Saliency‑Evaluation‑Toolbox.