Skip to content
main
Switch branches/tags
Go to file
Code

Latest commit

 

Git stats

Files

Permalink
Failed to load latest commit information.
Type
Name
Latest commit message
Commit time
 
 
 
 

README.md

SMAC: Learning Selective Mutual Attention and Contrast for RGB-D Saliency Detection

paper link: https://arxiv.org/abs/2010.05537

Citing our work

If you think our work is helful, please cite

@misc{liu2020learning,
      title={Learning Selective Mutual Attention and Contrast for RGB-D Saliency Detection}, 
      author={Nian Liu and Ni Zhang and Ling Shao and Junwei Han},
      year={2020},
      eprint={2010.05537},
      archivePrefix={arXiv},
      primaryClass={cs.CV}
}

The Proposed RGB-D Salient Object Detection Dataset

ReDWeb-S

We construct a new large-scale challenging dataset ReDWeb-S and it has totally 3179 images with various real-world scenes and high-quality depth maps. We split the dataset into a training set with 2179 RGB-D image pairs and a testing set with the remaining 1000 image pairs.

avatar

The proposed dataset link can be found here. [baidu pan fetch code: rp8b | Google drive]

Dataset Statistics and Comparisons

We analyze the proposed ReDWeb-S datset from several statistical aspects and also conduct a comparison between ReDWeb-S and other existing RGB-D SOD datasets.

avatar

avatar Fig.1. Top 60% scene and object category distributions of our proposed ReDWeb-S dataset.

avatar Fig.2. Comparison of nine RGB-D SOD dataset in terms of the distributions of global contrast and interior contrast.

avatar Fig.3. Comparsion of the average annotation maps for nine RGB-D SOD benchmark datasets.

avatar

Fig.4. Comparsion of the distribution of object size for nine RGB-D SOD benchmark datasets.

SOTA Results on our proposed dataset

We provide other SOTA RGB-D methods' results and scores on our proposed dataset. You can directly download all results [here lfa6].

No. Pub. Name Title Download
01 CVPR2020 S2MA Learning Selective Self-Mutual Attention for RGB-D Saliency Detection results, g0pgx
02 CVPR2020 JL-DCF JL-DCF: Joint Learning and Densely-Cooperative Fusion Framework for RGB-D Salient Object Detection results, xh9p
03 CVPR2020 UCNet UC-Net: Uncertainty Inspired RGB-D Saliency Detection via Conditional Variational Autoencoders results, 6o93
04 CVPR2020 A2dele A2dele: Adaptive and Attentive Depth Distiller for Efficient RGB-D Salient Object Detection results, swv5
05 CVPR2020 SSF-RGBD Select, Supplement and Focus for RGB-D Saliency Detection results, oshl
06 TIP2020 DisenFusion RGBD Salient Object Detection via Disentangled Cross-Modal Fusion results, h3hc
07 TNNLS2020 D3Net D3Net:Rethinking RGB-D Salient Object Detection: Models, Datasets, and Large-Scale Benchmarks results, tetn
08 ICCV2019 DMRA Depth-induced multi-scale recurrent attention network for saliency detection results, kqq4
09 CVPR2019 CPFP Depth-induced multi-scale recurrent attention network for saliency detection results, 0v2c
10 TIP2019 TANet Three-stream attention-aware network for RGB-D salient object detection results, hsy9
11 CVPR2018 PCF Progressively Complementarity-Aware Fusion Network for RGB-D Salient Object Detection results, qzhm
12 PR2019 MMCI Multi-modal fusion network with multiscale multi-path and cross-modal interactions for RGB-D salient object detection results, c90m
13 TCyb2017 CTMF CNNs-based RGB-D saliency detection via cross-view transfer and multiview fusion results, i0zb
14 Access2019 AFNet Adaptive fusion for rgb-d salient object detection results, 54zc
15 TIP2017 DF Rgbd salient object detection via deep fusion results, d7sc
16 ICME2016 SE Salient object detection for rgb-d image via saliency evolution results, h10s
17 SPL2016 DCMC Saliency detection for stereoscopic images based on depth confidence analysis and multiple cues fusion results, 18po
18 CVPR2016 LBE Local background enclosure for rgb-d salient object detection results, iiz5
Methods S-measure maxF E-measure MAE
S2MA 0.711 0.696 0.781 0.139
JL-DCF 0.734 0.727 0.805 0.128
UCNet 0.713 0.71 0.794 0.13
A2dele 0.641 0.603 0.672 0.16
SSF-RGBD 0.595 0.558 0.71 0.189
DisenFusion 0.675 0.658 0.76 0.16
D3Net 0.689 0.673 0.768 0.149
DMRA 0.592 0.579 0.721 0.188
CPFP 0.685 0.645 0.744 0.142
TANet 0.656 0.623 0.741 0.165
PCF 0.655 0.627 0.743 0.166
MMCI 0.660 0.641 0.754 0.176
CTMF 0.641 0.607 0.739 0.204
AFNet 0.546 0.549 0.693 0.213
DF 0.595 0.579 0.683 0.233
SE 0.435 0.393 0.587 0.283
DCMC 0.427 0.348 0.549 0.313
LBE 0.637 0.629 0.73 0.253

Acknowledgement

We thank all annotators for helping us constructing the proposed dataset. Our proposed dataset is based on the ReDWeb dataset, which is a state-of-the-art dataset proposed for monocular image depth estimation. We also thank the authors for providing the ReDWeb dataset.

Contact

If you have any questions, please feel free to contact me. (nnizhang.1995@gmail.com)

About

ReDWeb-S: a large-scale challenging dataset for RGB-D Salient Object Detection.

Resources

Releases

No releases published

Packages

No packages published

Languages