Awesome

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

arXiv version: https://arxiv.org/abs/2010.05537

Citing our work

If you think our work is helpful, please cite

@article{liu2021learning,
  title={Learning Selective Mutual Attention and Contrast for RGB-D Saliency Detection},
  author={Liu, Nian and Zhang, Ni and Shao, Ling and Han, Junwei},
  journal={IEEE Transactions on Pattern Analysis and Machine Intelligence},
  year={2021}
}

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 ov08].

No.	Pub.	Name	Title	Download
00	TIP2023	Caver	Caver: Cross-modal view-mixed transformer for bi-modal salient object detection	results, 2kfm
01	TCSVT2022	HRTransNet	HRTransNet: HRFormer-Driven Two-Modality Salient Object Detection	results, azjb
02	TCSVT2021	SwinNet	SwinNet: Swin Transformer Drives Edge-Aware RGB-D and RGB-T Salient Object Detection	results, zf9s
03	ICCV2021	CMINet	RGB-D Saliency Detection via Cascaded Mutual Information Minimization	results, maav
04	ICCV2021	VST	Visual Saliency Transformer	results, rkq9
05	ICCV2021	SPNet	Specificity-preserving RGB-D Saliency Detection	results, wwup
06	CVPR2021	DCF	Calibrated RGB-D Salient Object Detection	results, 3kn9
07	ECCV2020	PGAR	Progressively Guided Alternate Refinement Network for RGB-D Salient Object Detection	results, mwtr
08	ECCV2020	HDFNet	Hierarchical Dynamic Filtering Network for RGB-D Salient Object Detection	results, b98z
09	ECCV2020	DANet	A Single Stream Network for Robust and Real-time RGB-D Salient Object Detection	results, 1luj
10	ECCV2020	CoNet	Accurate RGB-D Salient Object Detection via Collaborative Learning	results, bqq6
11	ECCV2020	CMWNet	Cross-Modal Weighting Network for RGB-D Salient Object Detection	results, ztv9
12	ECCV2020	cmMS	RGB-D Salient Object Detection with Cross-Modality Modulation and Selection	results, kwe5
13	ECCV2020	BBS-Net	BBS-Net: RGB-D Salient Object Detection with a Bifurcated Backbone Strategy Network	results, ya5v
14	ECCV2020	ATSA	Asymmetric Two-Stream Architecture for Accurate RGB-D Saliency Detection	results, k750
15	CVPR2020	S2MA	Learning Selective Self-Mutual Attention for RGB-D Saliency Detection	results, g0pgx
16	CVPR2020	JL-DCF	JL-DCF: Joint Learning and Densely-Cooperative Fusion Framework for RGB-D Salient Object Detection	results, xh9p
17	CVPR2020	UCNet	UC-Net: Uncertainty Inspired RGB-D Saliency Detection via Conditional Variational Autoencoders	results, 6o93
18	CVPR2020	A2dele	A2dele: Adaptive and Attentive Depth Distiller for Efficient RGB-D Salient Object Detection	results, swv5
19	CVPR2020	SSF-RGBD	Select, Supplement and Focus for RGB-D Saliency Detection	results, oshl
20	TIP2020	DisenFusion	RGBD Salient Object Detection via Disentangled Cross-Modal Fusion	results, h3hc
21	TNNLS2020	D3Net	D3Net:Rethinking RGB-D Salient Object Detection: Models, Datasets, and Large-Scale Benchmarks	results, tetn
22	ICCV2019	DMRA	Depth-induced multi-scale recurrent attention network for saliency detection	results, kqq4
23	CVPR2019	CPFP	Contrast Prior and Fluid Pyramid Integration for RGBD Salient Object Detection	results, 0v2c
24	TIP2019	TANet	Three-stream attention-aware network for RGB-D salient object detection	results, hsy9
25	CVPR2018	PCF	Progressively Complementarity-Aware Fusion Network for RGB-D Salient Object Detection	results, qzhm
26	PR2019	MMCI	Multi-modal fusion network with multiscale multi-path and cross-modal interactions for RGB-D salient object detection	results, c90m
27	TCyb2017	CTMF	CNNs-based RGB-D saliency detection via cross-view transfer and multiview fusion	results, i0zb
28	Access2019	AFNet	Adaptive fusion for rgb-d salient object detection	results, 54zc
29	TIP2017	DF	Rgbd salient object detection via deep fusion	results, d7sc
30	ICME2016	SE	Salient object detection for rgb-d image via saliency evolution	results, h10s
31	SPL2016	DCMC	Saliency detection for stereoscopic images based on depth confidence analysis and multiple cues fusion	results, 18po
32	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)