Pith. sign in

REVIEW 3 major objections 5 minor 70 references

Lightweight RGB-D Salient Object Detection from a Speed-Accuracy Tradeoff Perspective

T0 review · 3 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A 5.2-million-parameter RGB-D salient object detector, SATNet, beats heavyweight CNN models on five benchmarks while running at 415 frames per second.

desk verdict The advertised 5.2M/415FPS describes only the downstream network, not the Depth Anything model that generates the pseudo depth, so the central speed-accuracy claim is not measured for the full system. read the letter →

arxiv 2505.04758 v1 pith:MRLAMLOE submitted 2025-05-07 cs.CV

Songsong Duan , Xi Yang , Nannan Wang , Xinbo Gao This is my paper · ORCID
classification cs.CV
keywords RGB-DsalientobjectdetectionlightweightnetworkpseudodepthAnythingModeldecoupledattentiondualinformationrepresentationspeed-accuracytradeoffreal-timeinference
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper claims that a lightweight RGB-D salient object detector can match or exceed the accuracy of much larger models if the depth input, fusion mechanism, and feature representation are redesigned for the lightweight regime. SATNet replaces noisy raw depth maps with pseudo depth maps from a monocular depth foundation model, uses a Decoupled Attention Module for cross-modal fusion, and expands the limited feature space of a MobileNet-v2 backbone with a Dual Information Representation Module that models texture and saliency separately. On five public benchmarks, the authors report that SATNet surpasses the best lightweight methods and several heavyweight CNN methods with 5.2 million parameters and 415 FPS. The claim matters because it suggests that the usual accuracy cost of small, real-time models can be substantially reduced by addressing data quality and feature space rather than simply adding parameters.

What carries the argument

The load-bearing machinery is a set of three lightweight modules built on a pair of MobileNet-v2 encoders. DAM (Decoupled Attention Module) pools RGB and depth features into horizontal and vertical vectors, projects them through fully connected layers into dual-view attention weights, and then cross-multiplies modality-specific spatial heatmaps so each branch is enhanced by the other's saliency evidence. DIRM (Dual Information Representation Module) creates two inverted pyramids, a texture pyramid supervised by edge ground truth and a saliency pyramid supervised by saliency ground truth, with two prediction heads that let gradients flow back through both; this is what expands the feature space without many extra parameters. DFAM (Dual Feature Aggregation Module) combines texture and saliency features and applies three branches of asymmetric and dilated depthwise separable convolutions to gather multi-scale receptive-field cues. The Depth Anything Model, used before training and testing, converts poor raw depth inputs into pseudo depth maps, which the paper argues removes the cross-modal inconsistency that hurts lightweight models.

What would settle it

Measure the full test-time pipeline, including the Depth Anything forward pass on each RGB image, on the same GPU used to report 415 FPS; if the end-to-end throughput is materially lower, the headline speed does not describe the deployable system.

Watch

Extended reading notes

Core claim

The paper's central claim is that the accuracy gap between lightweight and heavyweight RGB-D salient object detection is not primarily a parameter-count problem but a combination of poor depth quality, attention mechanisms unsuited to small models, and constrained feature representation. SATNet addresses each with a targeted module: Depth Anything pseudo depth maps supply smoother and more reliable geometric cues than raw depth; DAM decouples features into horizontal and vertical vectors and exchanges spatial heatmaps across modalities; DIRM builds texture and saliency feature pyramids with separate prediction heads; and DFAM merges those features using asymmetric and dilated depthwise separable convolutions that enlarge receptive fields without adding parameters. As reported on NLPR, NJU2K, SIP, STERE, and RGBD135, SATNet outperforms the convolutional heavyweight baselines and all six lightweight baselines with 5.2 M parameters, 1.5 G FLOPs, and 415 FPS.

Load-bearing premise

The reported speed and size figures assume that generating the pseudo depth map is free or excluded, and that pseudo depth is an acceptable replacement for raw depth when comparing with raw-depth baselines.

Editorial extensions

If this is right

  • If the reported numbers hold, lightweight RGB-D salient object detection can match heavyweight CNN accuracy, making saliency detection practical on phones and embedded devices.
  • The pseudo-depth strategy transfers to other architectures: re-training MobileSal and AirSOD with pseudo depth improved their MAE on SIP, so depth-quality substitution is a general recipe rather than a SATNet-specific trick.
  • The efficiency analysis, in which DAM and DIRM account for only 14.5% of parameters and 5.4% of runtime, implies that most of the accuracy gain comes from modules that are cheap to run.
  • With a Swin-tiny backbone, SATNet stays competitive with transformer-based RGB-D methods at a fraction of their FLOPs, so the design generalizes beyond CNN backbones.
  • The same network, with EfficientNet or Swin-tiny encoders, is competitive on polyp segmentation and RGB-T salient object detection, indicating the speed-accuracy tradeoff carries over to adjacent foreground segmentation tasks.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Beyond the paper's numbers, an end-to-end deployment that generates pseudo depth on the fly would add the Depth Anything forward pass cost, which the 415 FPS figure does not include.
  • The pseudo-depth advantage is likely strongest when raw depth is noisy and weakest when monocular depth estimation fails, such as with transparent or textureless objects; the paper does not test those failure regimes.
  • DAM's horizontal and vertical decoupling is a low-rank attention proxy that could transfer to other bi-modal fusion tasks, such as RGB-thermal saliency detection, where the same speed-accuracy pressure exists.
  • The separate texture and saliency heads make SATNet a form of multi-task learner, suggesting that edge labels could be distilled or omitted in future variants to reduce annotation cost.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The paper proposes SATNet, a lightweight RGB-D salient object detection network built on two MobileNet-v2 encoders, a Decoupled Attention Module (DAM), a Dual Information Representation Module (DIRM) with texture and saliency branches, and a decoder containing Dual Feature Aggregation Modules (DFAM). The Depth Anything Model is used to replace low-quality dataset depth maps with pseudo depth maps for both training and testing. The authors report strong accuracy on five RGB-D SOD benchmarks and claim a lightweight framework with 5.2 M parameters and 415 FPS, along with extensive ablations, failure cases, and extensions to polyp segmentation and RGB-T SOD.

Significance. If the reported efficiency numbers described the actual deployed system, the paper would offer a genuinely useful speed-accuracy tradeoff analysis: the module designs are concrete, the ablations are extensive, and the released code is a real strength. The DAM comparison against standard attention mechanisms and the DFAM comparison against ASPP/DenseASPP are informative, and Table V provides a partial control for the pseudo-depth input. However, the central efficiency claim is not measured for the full inference pipeline because the Depth Anything estimator is excluded from parameter, FLOPs, and FPS counts, and the accuracy comparison against raw-depth baselines is not controlled for the depth input source. These issues affect the headline contribution and require substantial revision.

major comments (3)
  1. [Abstract; III-A; Tables I and III] The headline claims of 5.2 M parameters and 415 FPS describe only the downstream SATNet, not the full system. In Fig. 2 and III-A, the Depth Anything Model is applied to every RGB input to produce the pseudo depth map consumed by the depth encoder, and the final model is trained and tested on pseudo depth (Table IV, case d). Without this model, the depth encoder has no input at test time. Table III lists per-stage time and parameters only for Encoder, DAM, DIRM, and Decoder, and Table I does not include the Depth Anything Model in Params, FLOPs, or FPS. The smallest Depth Anything variants add tens of millions of parameters and substantial per-image latency, so the stated 5.2 M/415 FPS cannot be reproduced for the system whose speed-accuracy tradeoff is the paper's title and central contribution. Please report full-system cost including the depth estimator, or explicitly reposition the claims as applying only to the downstream network and state the full-system numbers in the abstract and Table I.
  2. [IV-D, Table I; IV-E1, Table V] The accuracy comparison against SOTA baselines is not on equal footing. SATNet is trained and evaluated with pseudo depth from Depth Anything, while the baselines in Table I are evaluated with the low-quality dataset depth (or their own depth inputs). Table V shows that retraining MobileSal and AirSOD on pseudo depth improves their SIP MAE from 0.057 to 0.052 and from 0.060 to 0.054, respectively, so a nontrivial part of SATNet's advantage over these lightweight methods may come from the input representation rather than the proposed architecture. The claim in IV-D that SATNet 'excels SOTA CNN-based heavyweight models' is not supportable until the baselines are given the same pseudo-depth input or the comparison is otherwise controlled. At minimum, all lightweight baselines in Table I should be retrained with pseudo depth, and the speed comparison should include the cost of generating pseudo depth.
  3. [IV-E4 and IV-E5; Tables IX and X] The hyperparameters psi_ef and the DFAM kernel/dilation combinations are selected by evaluating on the SIP and NLPR test sets, which are the same datasets used for the final headline numbers in Tables I, VI, VII, and VIII. The paper does not describe a validation split or cross-validation for these choices. This is test-set tuning: values such as psi_ef=32 and CK&DR=(3,5,7) are chosen to optimize the reported metrics, so the final results are not an independent evaluation of the configuration. Please introduce a held-out validation protocol for hyperparameter selection and report final results on a genuine test protocol, or clearly acknowledge this tuning and temper the corresponding claims.
minor comments (5)
  1. [IV-D1] The sentence 'Ours vs AirSOD: 365 vs 415' reverses the order; Table I lists AirSOD at 365 FPS and SATNet at 415 FPS.
  2. [Table VI] The row 'Ours-Conv3' contains an extra 0.045 and appears to omit the NLPR S-measure value; please correct the table.
  3. [II-A] The sentence 'we focus on lightweight RGB SOD models' should presumably read 'lightweight RGB-D SOD models'.
  4. [III-C and Table VII] The module is introduced as Global Semantic Refinement (GSR), but the ablation text and Table VII also use 'GRS'; please standardize the abbreviation.
  5. [Abstract] The phrase 'to optimal its parameters' should be 'to optimize its parameters'.

Circularity Check

1 steps flagged · score 5.0 of 10

Hyperparameters are selected on the SIP/NLPR test sets, so the reported SIP/NLPR accuracy is the selection criterion rather than an independent prediction.

  1. fitted input called prediction [Section IV-E.4 and IV-E.5 (Tables IX and X)]
    "To analyze the role of different convolution kernels and dilated rates, we set different kernel sizes and rates in Table IX. The best performance is same as our SATNet in manuscript. ... To explore the optimal setting, we formulate different variants with different ψef (16, 64, and 128), denoted as EF16, EF64, and EF128, where results of these variants are presented in Table X. ... Therefore, the efficiency factor is set to 32 is a suitable choice."

    The final SATNet configuration is chosen by taking the best-performing variant on the SIP and NLPR test sets: Table X selects ψef=32 after comparing EF16/EF64/EF128, and Table IX selects kernel/dilation rates (3,5,7) because that is the best row on those same test sets. The 'Ours' rows in those tables are then reported as SATNet's headline accuracy on SIP and NLPR in Table I. Thus the claim that SATNet outperforms its variants on SIP/NLPR is true by construction: the model is defined as the argmax of those metrics on those test sets, and the reported numbers are the selection objective itself, not an independent evaluation. The remaining three benchmarks were not used for hyperparameter selection, so the circularity is partial.

full rationale

The paper's modular claims (DAM, DIRM, DFAM) are evaluated against external benchmarks and standard ablations; there is no derivation-level equivalence in which an output equation reduces to an input by definition. The Depth Anything Model is an external module, not derived here, and its omission from the parameter/FPS count is a reporting-fairness caveat rather than a circular derivation. The only concrete circularity is the model-selection procedure: ψef and the kernel/dilation rates are tuned by maximizing performance on the SIP and NLPR test sets, and the resulting selected configuration is then presented as 'Ours' with those same test-set numbers as evidence. This is test-set double dipping and inflates the reported improvements on those two datasets, which are also used in the headline comparisons. Because the architecture itself has independent content and the other three datasets are not part of the selection criterion, the overall circularity is partial rather than total.

Assumptions & free parameters 4 free parameters · 3 assumptions · 0 invented entities

The paper relies on standard benchmark assumptions and a domain assumption that pseudo depth is a valid substitute. The main free parameters are hyperparameters selected on the test sets used for final evaluation, which adds mild fitting risk. No new physical entities are introduced.

free parameters (4)
  • efficiency factor psi_ef = 32
    Channel width after fusion; chosen via ablation on SIP and NLPR test sets (Table X).
  • DFAM kernel/dilation rates = (3,5,7)
    Convolution kernel sizes and dilated rates in DFAM; selected via ablation on SIP and NLPR (Table IX).
  • DAM spatial kernel = 7x7
    Conv7 in DAM cross-interaction; 7x7 vs 3x3 variant in Table VI.
  • loss weights = 1, 1, 1
    BCE, IOU, and SSIM losses combined with equal weights (Eq. 13); no tuning reported.
assumptions (3)
  • domain assumption Pseudo depth maps from Depth Anything are a valid and better surrogate for sensor depth in RGB-D SOD.
    Used throughout; motivated in Section I and III-A, tested only internally.
  • domain assumption Standard dataset splits (NLPR 700/300, NJU2K 1485/500, etc.) and metrics are accepted without independent verification.
    Section IV-A/B; these benchmarks are community standards.
  • domain assumption Pretrained MobileNetV2 and Depth Anything transfer to SOD without overfitting to the test images.
    Assumed in Section IV-C; no explicit test for data leakage.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Lightweight RGB-D Salient Object Detection from a Speed-Accuracy Tradeoff Perspective." pith.science (2026). https://pith.science/paper/MRLAMLOE

@misc{pith2026250504758,
  author       = {Pith},
  title        = {Pith review of: Lightweight RGB-D Salient Object Detection from a Speed-Accuracy Tradeoff Perspective},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/MRLAMLOE}},
  note         = {Machine review of arXiv:2505.04758}
}
read the original abstract

Current RGB-D methods usually leverage large-scale backbones to improve accuracy but sacrifice efficiency. Meanwhile, several existing lightweight methods are difficult to achieve high-precision performance. To balance the efficiency and performance, we propose a Speed-Accuracy Tradeoff Network (SATNet) for Lightweight RGB-D SOD from three fundamental perspectives: depth quality, modality fusion, and feature representation. Concerning depth quality, we introduce the Depth Anything Model to generate high-quality depth maps,which effectively alleviates the multi-modal gaps in the current datasets. For modality fusion, we propose a Decoupled Attention Module (DAM) to explore the consistency within and between modalities. Here, the multi-modal features are decoupled into dual-view feature vectors to project discriminable information of feature maps. For feature representation, we develop a Dual Information Representation Module (DIRM) with a bi-directional inverted framework to enlarge the limited feature space generated by the lightweight backbones. DIRM models texture features and saliency features to enrich feature space, and employ two-way prediction heads to optimal its parameters through a bi-directional backpropagation. Finally, we design a Dual Feature Aggregation Module (DFAM) in the decoder to aggregate texture and saliency features. Extensive experiments on five public RGB-D SOD datasets indicate that the proposed SATNet excels state-of-the-art (SOTA) CNN-based heavyweight models and achieves a lightweight framework with 5.2 M parameters and 415 FPS.

Figures

Figures reproduced from arXiv: 2505.04758 by the authors.

Figure 1
Figure 1. Illustration of the motivation for our SATNet. (a) is [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Framework of proposed SATNet. It consists of four parts, including RGB and Depth Encoders, Decoupled Attention [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 5
Figure 5. The feature visualization of the proposed DIRM, [PITH_FULL_IMAGE:figures/full_fig_p005_5.png] view at source ↗
Figures from the paper (7 more)
Figure 4
Figure 4. Figure 4: The visualization of the proposed DAM and other [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 6
Figure 6. Figure 6: The illustration of the proposed DFAM. features of TFP, where the LTR module effectively learns the local texture prior PLT . Specifically, we firstly integrate the multi-modality fused features f i f via connected layer and 1×1 convolution operation. The total process…
Figure 7
Figure 7. Figure 7: Precision-Recall curves of our model and other RGB-D SOD methods on NLPR, NJU2K, SIP, STERE, and RGBD135 [PITH_FULL_IMAGE:figures/full_fig_p008_7.png]
Figure 8
Figure 8. Figure 8: Qualitative comparison of some SOTA RGB-D methods and our model. [PITH_FULL_IMAGE:figures/full_fig_p009_8.png]
Figure 9
Figure 9. Figure 9: Visual comparison of edge maps, including our SAT [PITH_FULL_IMAGE:figures/full_fig_p009_9.png]
Figure 10
Figure 10. Figure 10: The visual comparison of surface normals (SN). [PITH_FULL_IMAGE:figures/full_fig_p010_10.png]
Figure 11
Figure 11. Figure 11: Some failure cases of our SATNet, LSNet [ [PITH_FULL_IMAGE:figures/full_fig_p012_11.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

70 extracted references · 55 canonical work pages

  1. [1]

    Sketch-based image retrieval by salient contour reinforcement,

    Y . Zhang, X. Qian, X. Tan, J. Han, and Y . Tang, “Sketch-based image retrieval by salient contour reinforcement,” IEEE Transactions on Multimedia, vol. 18, no. 8, pp. 1604–1615, 2016

    work page 2016

  2. [2]

    Non- rigid object tracking via deep multi-scale spatial-temporal discriminative saliency maps,

    P. Zhang, W. Liu, D. Wang, Y . Lei, H. Wang, and H. Lu, “Non- rigid object tracking via deep multi-scale spatial-temporal discriminative saliency maps,” Pattern Recognition, vol. 100, p. 107130, 2020

    work page 2020

  3. [3]

    Pranet: Parallel reverse attention network for polyp segmentation,

    D.-P. Fan, G.-P. Ji, T. Zhou, G. Chen, H. Fu, J. Shen, and L. Shao, “Pranet: Parallel reverse attention network for polyp segmentation,” in Proceedings of the International Conference on Medical Image Computing and Computer-Assisted Intervention , 2020, pp. 263–273

    work page 2020

  4. [4]

    Non-salient region object mining for weakly supervised semantic segmentation,

    Y . Yao, T. Chen, G.-S. Xie, C. Zhang, F. Shen, Q. Wu, Z. Tang, and J. Zhang, “Non-salient region object mining for weakly supervised semantic segmentation,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2021, pp. 2623–2632

    2021

  5. [5]

    Depth anything: Unleashing the power of large-scale unlabeled data,

    L. Yang, B. Kang, Z. Huang, X. Xu, J. Feng, and H. Zhao, “Depth anything: Unleashing the power of large-scale unlabeled data,” arXiv preprint arXiv:2401.10891, 2024

    arXiv 2024

  6. [6]

    Squeeze-and-excitation networks,

    J. Hu, L. Shen, and G. Sun, “Squeeze-and-excitation networks,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2018, pp. 7132–7141

    2018

  7. [7]

    Cbam: Convolutional block attention module,

    S. Woo, J. Park, J.-Y . Lee, and I. S. Kweon, “Cbam: Convolutional block attention module,” in Proceedings of the European Conference on Computer Vision, 2018, pp. 3–19

    2018

  8. [8]

    Attention is all you need,

    A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, L. Kaiser, and I. Polosukhin, “Attention is all you need,” in Proceedings of the Advances in Neural Information Processing Systems , vol. 30, 2017

    work page 2017

Show all 70 references
  1. [9]

    Hi- danet: Rgb-d salient object detection via hierarchical depth awareness,

    Z. Wu, G. Allibert, F. Meriaudeau, C. Ma, and C. Demonceaux, “Hi- danet: Rgb-d salient object detection via hierarchical depth awareness,” IEEE Transactions on Image Processing, vol. 32, pp. 2160–2173, 2023

    2023

  2. [10]

    Dgfnet: Depth-guided cross- modality fusion network for rgb-d salient object detection,

    F. Xiao, Z. Pu, J. Chen, and X. Gao, “Dgfnet: Depth-guided cross- modality fusion network for rgb-d salient object detection,” IEEE Transactions on Multimedia , vol. 26, pp. 2648–2658, 2024

    2024

  3. [11]

    Learning implicit class knowledge for rgb-d co-salient object detection with transformers,

    N. Zhang, J. Han, and N. Liu, “Learning implicit class knowledge for rgb-d co-salient object detection with transformers,” IEEE Transactions on Image Processing , vol. 31, pp. 4556–4570, 2022

    2022

  4. [13]

    Transformer fusion and pixel- level contrastive learning for rgb-d salient object detection,

    J. Wu, F. Hao, W. Liang, and J. Xu, “Transformer fusion and pixel- level contrastive learning for rgb-d salient object detection,” IEEE Transactions on Multimedia , 2023

    2023

  5. [14]

    A simple pooling- based design for real-time salient object detection,

    J.-J. Liu, Q. Hou, M.-M. Cheng, J. Feng, and J. Jiang, “A simple pooling- based design for real-time salient object detection,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2019, pp. 3917–3926

    2019

  6. [15]

    Larnet: Towards lightweight, accurate and real-time salient object detection,

    Z. Wang, Y . Zhang, Y . Liu, C. Qin, S. A. Coleman, and D. Kerr, “Larnet: Towards lightweight, accurate and real-time salient object detection,” IEEE Transactions on Multimedia , 2023

    2023

  7. [16]

    Depthwise nonlocal module for fast salient object detection using a single thread,

    H. Li, G. Li, B. Yang, G. Chen, L. Lin, and Y . Yu, “Depthwise nonlocal module for fast salient object detection using a single thread,” IEEE Transactions on Cybernetics , vol. 51, no. 12, pp. 6188–6199, 2020

    2020

  8. [17]

    Feature calibrating and fusing network for rgb-d salient object detection,

    Q. Zhang, Q. Qin, Y . Yang, Q. Jiao, and J. Han, “Feature calibrating and fusing network for rgb-d salient object detection,” IEEE Transactions on Circuits and Systems for Video Technology , 2024

    2024

  9. [18]

    Disentangled cross- modal transformer for rgb-d salient object detection and beyond,

    H. Chen, F. Shen, D. Ding, Y . Deng, and C. Li, “Disentangled cross- modal transformer for rgb-d salient object detection and beyond,” IEEE Transactions on Image Processing , vol. 33, pp. 1699–1709, 2024

    2024

  10. [19]

    3-d convolutional neural networks for rgb-d salient object detection and beyond,

    Q. Chen, Z. Zhang, Y . Lu, K. Fu, and Q. Zhao, “3-d convolutional neural networks for rgb-d salient object detection and beyond,” IEEE Transactions on Neural Networks and Learning Systems , vol. 35, no. 3, pp. 4309–4323, 2022

    2022

  11. [20]

    Robust perception and precise segmentation for scribble- supervised rgb-d saliency detection,

    L. Li, J. Han, N. Liu, S. Khan, H. Cholakkal, R. M. Anwer, and F. S. Khan, “Robust perception and precise segmentation for scribble- supervised rgb-d saliency detection,” IEEE Transactions on Pattern Analysis and Machine Intelligence , vol. 46, no. 1, pp. 479–496, 2024

    2024

  12. [21]

    Point-aware interaction and cnn-induced refinement network for rgb- d salient object detection,

    R. Cong, H. Liu, C. Zhang, W. Zhang, F. Zheng, R. Song, and S. Kwong, “Point-aware interaction and cnn-induced refinement network for rgb- d salient object detection,” in Proceedings of the ACM International Conference on Multimedia , 2023, pp. 406–416

    2023

  13. [22]

    Modality-induced transfer-fusion network for rgb-d and rgb-t salient object detection,

    G. Chen, F. Shao, X. Chai, H. Chen, Q. Jiang, X. Meng, and Y .-S. Ho, “Modality-induced transfer-fusion network for rgb-d and rgb-t salient object detection,” IEEE Transactions on Circuits and Systems for Video Technology, vol. 33, no. 4, pp. 1787–1801, 2023

    2023

  14. [23]

    Hfmdnet: Hierarchical fusion and multi-level decoder network for rgb- d salient object detection,

    Y . Luo, F. Shao, Z. Xie, H. Wang, H. Chen, B. Mu, and Q. Jiang, “Hfmdnet: Hierarchical fusion and multi-level decoder network for rgb- d salient object detection,” IEEE Transactions on Instrumentation and Measurement, 2024

    2024

  15. [24]

    A2dele: Adaptive and attentive depth distiller for efficient rgb-d salient object detection,

    Y . Piao, Z. Rong, M. Zhang, W. Ren, and H. Lu, “A2dele: Adaptive and attentive depth distiller for efficient rgb-d salient object detection,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2020, pp. 9060–9069

    2020

  16. [25]

    Salient object detection in rgb-d videos,

    A. Mou, Y . Lu, J. He, D. Min, K. Fu, and Q. Zhao, “Salient object detection in rgb-d videos,” IEEE Transactions on Image Processing , 2024

    2024

  17. [26]

    Lsnet: Lightweight spatial boosting network for detecting salient objects in rgb-thermal images,

    W. Zhou, Y . Zhu, J. Lei, R. Yang, and L. Yu, “Lsnet: Lightweight spatial boosting network for detecting salient objects in rgb-thermal images,” IEEE Transactions on Image Processing, vol. 32, pp. 1329–1340, 2023. 14

    2023

  18. [27]

    Airsod: A lightweight network for rgb-d salient object detection,

    Z. Zeng, H. Liu, F. Chen, and X. Tan, “Airsod: A lightweight network for rgb-d salient object detection,” IEEE Transactions on Circuits and Systems for Video Technology, vol. 34, no. 3, pp. 1656–1669, 2024

    2024

  19. [28]

    Mobilesal: Extremely efficient rgb-d salient object detection,

    Y .-H. Wu, Y . Liu, J. Xu, J.-W. Bian, Y .-C. Gu, and M.-M. Cheng, “Mobilesal: Extremely efficient rgb-d salient object detection,” IEEE Transactions on Pattern Analysis and Machine Intelligence , vol. 44, no. 12, pp. 10 261–10 269, 2021

    2021

  20. [29]

    Moadnet: Mobile asymmetric dual-stream networks for real-time and lightweight rgb-d salient object detection,

    X. Jin, K. Yi, and J. Xu, “Moadnet: Mobile asymmetric dual-stream networks for real-time and lightweight rgb-d salient object detection,” IEEE Transactions on Circuits and Systems for Video Technology , vol. 32, no. 11, pp. 7632–7645, 2022

    2022

  21. [30]

    Mobilenets: Efficient convo- lutional neural networks for mobile vision applications,

    A. G. Howard, M. Zhu, B. Chen, D. Kalenichenko, W. Wang, T. Weyand, M. Andreetto, and H. Adam, “Mobilenets: Efficient convo- lutional neural networks for mobile vision applications,” arXiv preprint arXiv:1704.04861, 2017

    2017 arXiv

  22. [31]

    Deeplab: Semantic image segmentation with deep convolutional nets, atrous convolution, and fully connected crfs,

    L.-C. Chen, G. Papandreou, I. Kokkinos, K. Murphy, and A. L. Yuille, “Deeplab: Semantic image segmentation with deep convolutional nets, atrous convolution, and fully connected crfs,” IEEE Transactions on Pattern Analysis and Machine Intelligence , vol. 40, no. 4, pp. 834–848, 2017

    2017

  23. [32]

    Denseaspp for semantic segmentation in street scenes,

    M. Yang, K. Yu, C. Zhang, Z. Li, and K. Yang, “Denseaspp for semantic segmentation in street scenes,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2018, pp. 3684–3692

    2018

  24. [33]

    An image is worth 16x16 words: Transformers for image recognition at scale,

    A. Dosovitskiy, L. Beyer, A. Kolesnikov, D. Weissenborn, X. Zhai, T. Unterthiner, M. Dehghani, M. Minderer, G. Heigold, S. Gelly et al., “An image is worth 16x16 words: Transformers for image recognition at scale,” in Proceedings of the International Conference on Learning Rep...

    2021

  25. [34]

    Lanet: Local attention embedding to improve the semantic segmentation of remote sensing images,

    L. Ding, H. Tang, and L. Bruzzone, “Lanet: Local attention embedding to improve the semantic segmentation of remote sensing images,” IEEE Transactions on Geoscience and Remote Sensing , vol. 59, no. 1, pp. 426–435, 2021

    2021

  26. [35]

    Non-local neural net- works,

    X. Wang, R. Girshick, A. Gupta, and K. He, “Non-local neural net- works,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2018, pp. 7794–7803

    2018

  27. [36]

    A convnet for the 2020s,

    Z. Liu, H. Mao, C.-Y . Wu, C. Feichtenhofer, T. Darrell, and S. Xie, “A convnet for the 2020s,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2022, pp. 11 976–11 986

    2022

  28. [37]

    Swin transformer: Hierarchical vision transformer using shifted windows,

    Z. Liu, Y . Lin, Y . Cao, H. Hu, Y . Wei, Z. Zhang, S. Lin, and B. Guo, “Swin transformer: Hierarchical vision transformer using shifted windows,” in Proceedings of the IEEE/CVF International Conference on Computer Vision, 2021, pp. 10 012–10 022

    2021

  29. [38]

    Acnet: Strengthening the kernel skeletons for powerful cnn via asymmetric convolution blocks,

    X. Ding, Y . Guo, G. Ding, and J. Han, “Acnet: Strengthening the kernel skeletons for powerful cnn via asymmetric convolution blocks,” in Proceedings of the IEEE/CVF International Conference on Computer Vision, 2019, pp. 1911–1920

    2019

  30. [39]

    Dilated residual networks,

    F. Yu, V . Koltun, and T. Funkhouser, “Dilated residual networks,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2017, pp. 472–480

    2017

  31. [40]

    Rethinking rgb-d salient object detection: Models, data sets, and large-scale bench- marks,

    D.-P. Fan, Z. Lin, Z. Zhang, M. Zhu, and M.-M. Cheng, “Rethinking rgb-d salient object detection: Models, data sets, and large-scale bench- marks,” IEEE Transactions on Neural Networks and Learning Systems , vol. 32, no. 5, pp. 2075–2089, 2020

    2020

  32. [41]

    Depth enhanced saliency detection method,

    Y . Cheng, H. Fu, X. Wei, J. Xiao, and X. Cao, “Depth enhanced saliency detection method,” in Proceedings of International Conference on Internet Multimedia Computing and Service , 2014, pp. 23–27

    2014

  33. [42]

    Leveraging stereopsis for saliency analysis,

    Y . Niu, Y . Geng, X. Li, and F. Liu, “Leveraging stereopsis for saliency analysis,” in Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2012, pp. 454–461

    2012

  34. [43]

    Rgbd salient object detection: a benchmark and algorithms,

    H. Peng, B. Li, W. Xiong, W. Hu, and R. Ji, “Rgbd salient object detection: a benchmark and algorithms,” in Proceedings of the European Conference on Computer Vision , 2014, pp. 92–109

    2014

  35. [44]

    Enhanced-alignment measure for binary foreground map evaluation,

    D.-P. Fan, C. Gong, Y . Cao, B. Ren, M.-M. Cheng, and A. Borji, “Enhanced-alignment measure for binary foreground map evaluation,” Proceedings of the International Joint Conference on Artificial Intelli- gence, pp. 698–704, 2018

    2018

  36. [45]

    Structure- measure: A new way to evaluate foreground maps,

    D.-P. Fan, M.-M. Cheng, Y . Liu, T. Li, and A. Borji, “Structure- measure: A new way to evaluate foreground maps,” in Proceedings of the IEEE/CVF International Conference on Computer Vision , 2017, pp. 4548–4557

    2017

  37. [46]

    Frequency-tuned salient region detection,

    R. Achanta, S. Hemami, F. Estrada, and S. Susstrunk, “Frequency-tuned salient region detection,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , 2009, pp. 1597–1604

    2009

  38. [47]

    How to evaluate foreground maps?

    R. Margolin, L. Zelnik-Manor, and A. Tal, “How to evaluate foreground maps?” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2014, pp. 248–255

    2014

  39. [48]

    Icnet: Information conversion network for rgb-d based salient object detection,

    G. Li, Z. Liu, and H. Ling, “Icnet: Information conversion network for rgb-d based salient object detection,” IEEE Transactions on Image Processing, vol. 29, pp. 4873–4884, 2020

    2020

  40. [49]

    Deep rgb-d saliency detection with depth-sensitive attention and automatic multi-modal fu- sion,

    P. Sun, W. Zhang, H. Wang, S. Li, and X. Li, “Deep rgb-d saliency detection with depth-sensitive attention and automatic multi-modal fu- sion,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2021, pp. 1407–1417

    2021

  41. [50]

    Bilateral attention network for rgb-d salient object detection,

    Z. Zhang, Z. Lin, J. Xu, W.-D. Jin, S.-P. Lu, and D.-P. Fan, “Bilateral attention network for rgb-d salient object detection,” IEEE Transactions on Image Processing , vol. 30, pp. 1949–1961, 2021

    1949

  42. [51]

    Siamese network for rgb-d salient object detection and beyond,

    K. Fu, D.-P. Fan, G.-P. Ji, Q. Zhao, J. Shen, and C. Zhu, “Siamese network for rgb-d salient object detection and beyond,” IEEE Transac- tions on Pattern Analysis and Machine Intelligence , vol. 44, no. 9, pp. 5541–5559, 2021

    2021

  43. [52]

    Depth- induced gap-reducing network for rgb-d salient object detection: an interaction, guidance and refinement approach,

    X. Cheng, X. Zheng, J. Pei, H. Tang, Z. Lyu, and C. Chen, “Depth- induced gap-reducing network for rgb-d salient object detection: an interaction, guidance and refinement approach,” IEEE Transactions on Multimedia, 2022

    2022

  44. [53]

    Multi-prior driven network for rgb-d salient object detection,

    X. Zhang, Y . Xu, T. Wang, and T. Liao, “Multi-prior driven network for rgb-d salient object detection,” IEEE Transactions on Circuits and Systems for Video Technology, 2023

    2023

  45. [54]

    Cross-modal hierarchical interaction network for rgb-d salient object detection,

    H. Bi, R. Wu, Z. Liu, H. Zhu, C. Zhang, and T.-Z. Xiang, “Cross-modal hierarchical interaction network for rgb-d salient object detection,” Pattern Recognition, vol. 136, p. 109194, 2023

    2023

  46. [55]

    Delving into cal- ibrated depth for accurate rgb-d salient object detection,

    J. Li, W. Ji, M. Zhang, Y . Piao, H. Lu, and L. Cheng, “Delving into cal- ibrated depth for accurate rgb-d salient object detection,” International Journal of Computer Vision , vol. 131, no. 4, pp. 855–876, 2023

    2023

  47. [56]

    Depth quality- inspired feature manipulation for efficient rgb-d salient object detection,

    W. Zhang, G.-P. Ji, Z. Wang, K. Fu, and Q. Zhao, “Depth quality- inspired feature manipulation for efficient rgb-d salient object detection,” in Proceedings of the ACM International Conference on Multimedia , 2021, pp. 731–740

    2021

  48. [57]

    Em- trans: Edge-aware multimodal transformer for rgb-d salient object de- tection,

    G. Chen, Q. Wang, B. Dong, R. Ma, N. Liu, H. Fu, and Y . Xia, “Em- trans: Edge-aware multimodal transformer for rgb-d salient object de- tection,” IEEE Transactions on Neural Networks and Learning Systems , 2024

    2024

  49. [58]

    Catnet: A cascaded and aggregated transformer network for rgb-d salient object detection,

    F. Sun, P. Ren, B. Yin, F. Wang, and H. Li, “Catnet: A cascaded and aggregated transformer network for rgb-d salient object detection,” IEEE Transactions on Multimedia , 2024

    2024

  50. [59]

    Tokens-to-token vit: Training vision transformers from scratch on imagenet,

    L. Yuan, Y . Chen, T. Wang, W. Yu, Y . Shi, Z. Jiang, F. E. H. Tay, J. Feng, and S. Yan, “Tokens-to-token vit: Training vision transformers from scratch on imagenet,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition . IEEE, 2021, pp. 3156– 3164

    2021

  51. [60]

    Saliency prototype for rgb-d and rgb- t salient object detection,

    Z. Zhang, J. Wang, and Y . Han, “Saliency prototype for rgb-d and rgb- t salient object detection,” in Proceedings of the ACM International Conference on Multimedia , 2023, pp. 3696–3705

    2023

  52. [61]

    Caver: Cross-modal view- mixed transformer for bi-modal salient object detection,

    Y . Pang, X. Zhao, L. Zhang, and H. Lu, “Caver: Cross-modal view- mixed transformer for bi-modal salient object detection,” IEEE Trans- actions on Image Processing , vol. 32, pp. 892–904, 2023

    2023

  53. [62]

    Glconet: Learning mul- tisource perception representation for camouflaged object detection,

    Y . Sun, H. Xuan, J. Yang, and L. Luo, “Glconet: Learning mul- tisource perception representation for camouflaged object detection,” IEEE Transactions on Neural Networks and Learning Systems , 2024

    2024

  54. [63]

    Boundary uncertainty aware network for automated polyp segmenta- tion,

    G. Yue, G. Zhuo, W. Yan, T. Zhou, C. Tang, P. Yang, and T. Wang, “Boundary uncertainty aware network for automated polyp segmenta- tion,” Neural Networks, vol. 170, pp. 390–404, 2024

    2024

  55. [64]

    Pranet: Parallel reverse attention network for polyp segmentation,

    D.-P. Fan, G.-P. Ji, T. Zhou, G. Chen, H. Fu, J. Shen, and L. Shao, “Pranet: Parallel reverse attention network for polyp segmentation,” in International Conference on Medical Image Computing and Computer Assisted Intervention, 2020, pp. 263–273

    2020

  56. [65]

    Polyp segmentation via semantic en- hanced perceptual network,

    T. Wang, X. Qi, and G. Yang, “Polyp segmentation via semantic en- hanced perceptual network,” IEEE Transactions on Circuits and Systems for Video Technology, 2024

    2024

  57. [66]

    Ctnet: Contrastive transformer network for polyp segmentation,

    B. Xiao, J. Hu, W. Li, C.-M. Pun, and X. Bi, “Ctnet: Contrastive transformer network for polyp segmentation,” IEEE Transactions on Cybernetics, 2024

    2024

  58. [67]

    Swinnet: Swin transformer drives edge-aware rgb-d and rgb-t salient object detection,

    Z. Liu, Y . Tan, Q. He, and Y . Xiao, “Swinnet: Swin transformer drives edge-aware rgb-d and rgb-t salient object detection,” IEEE Transactions on Circuits and Systems for Video Technology, vol. 32, no. 7, pp. 4486– 4497, 2021

    2021

  59. [68]

    Ecffnet: Effective and consistent feature fusion network for rgb-t salient object detection,

    W. Zhou, Q. Guo, J. Lei, L. Yu, and J.-N. Hwang, “Ecffnet: Effective and consistent feature fusion network for rgb-t salient object detection,” IEEE Transactions on Circuits and Systems for Video Technology , vol. 32, no. 3, pp. 1224–1235, 2021

    2021

  60. [69]

    Cgfnet: Cross-guided fusion network for rgb-t salient object detection,

    J. Wang, K. Song, Y . Bao, L. Huang, and Y . Yan, “Cgfnet: Cross-guided fusion network for rgb-t salient object detection,” IEEE Transactions on Circuits and Systems for Video Technology , vol. 32, no. 5, pp. 2949– 2961, 2021. 15

    2021

  61. [70]

    Tritransnet: Rgb-d salient object detection with a triplet transformer embedding network,

    Z. Liu, Y . Wang, Z. Tu, Y . Xiao, and B. Tang, “Tritransnet: Rgb-d salient object detection with a triplet transformer embedding network,” in Proceedings of the 29th ACM international conference on multimedia, 2021, pp. 4481–4490

    2021

  62. [71]

    Object segmentation by mining cross-modal semantics,

    Z. Wu, J. Wang, Z. Zhou, Z. An, Q. Jiang, C. Demonceaux, G. Sun, and R. Timofte, “Object segmentation by mining cross-modal semantics,” in Proceedings of the 31st ACM International Conference on Multimedia , 2023, pp. 3455–3464. Songsong Duan (Student Member, IEEE) received th...

    2023

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.