pith. sign in

arxiv: 2605.15535 · v1 · pith:6TQOLQGUnew · submitted 2026-05-15 · 💻 cs.CV

Learning Dynamic Structural Specialization for Underwater Salient Object Detection

Lin Hong , Chenhui Wang , Linan Deng , Yuning Cui , Yu Zhang , Xin Wang , Bojian Zhang , Wenqi Ren
show 2 more authors
Xingchen Yang Fumin Zhang
This is my paper

Pith reviewed 2026-05-19 14:20 UTC · model grok-4.3

classification 💻 cs.CV
keywords underwater salient object detectiondynamic structural specializationboundary-sensitive branchregion-coherent branchspatial coordinationcooperative structural supervisionsalient object detection
0
0 comments X p. Extension
pith:6TQOLQGU Add to your LaTeX paper What is a Pith Number? 
\usepackage{pith}
\pithnumber{6TQOLQGU}

Prints a linked pith:6TQOLQGU badge after your title and writes the identifier into PDF metadata. Compiles on arXiv with no extra files. Learn more

The pith

Dynamic structural specialization enhances underwater salient object detection by coordinating boundary and region features.

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

This paper presents DSS-USOD, a method for detecting salient objects in underwater images that suffer from degradations like poor visibility. The approach extracts a shared base representation and decomposes it into a boundary-sensitive branch for fine details and a region-coherent branch for structural consistency. A spatial coordination module then adjusts the influence of each branch based on the local image context to achieve better results. Cooperative structural supervision is used to encourage the branches to specialize effectively. Experiments demonstrate improved performance on standard benchmarks and successful real-world use on an underwater robot.

Core claim

The central discovery is that dynamically specializing a shared representation into boundary-sensitive and region-coherent structural features, coordinated by a spatial module according to local context, allows for more accurate localization, coherent regions, and precise boundaries in underwater salient object detection despite image degradations.

What carries the argument

dynamic structural specialization, which decomposes shared features into boundary-sensitive and region-coherent branches regulated by a spatial coordination module

Load-bearing premise

That decomposing the shared base representation into boundary-sensitive and region-coherent branches and regulating them with a spatial coordination module will correct inaccurate localization, fragmented regions, and coarse boundaries caused by underwater degradations.

What would settle it

Observing no improvement in boundary accuracy or region coherence when the spatial coordination or branch decomposition is removed in controlled experiments on degraded underwater images.

Figures

Figures reproduced from arXiv: 2605.15535 by Bojian Zhang, Chenhui Wang, Fumin Zhang, Linan Deng, Lin Hong, Wenqi Ren, Xingchen Yang, Xin Wang, Yuning Cui, Yu Zhang.

Figure 1
Figure 1. Figure 1: (a) Overview of dynamic structural specialization design in DSS [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Overall architecture of the proposed DSS-USOD. Given an underwater RGB image [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: PR curves and F-measure curves on the USOD10K and USOD. [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Qualitative visual comparison of DSS-USOD with 40 SOTA methods on USOD10K (first three rows) and USOD (last three rows). Owing to its [PITH_FULL_IMAGE:figures/full_fig_p008_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Trade-off between model complexity and performance. The horizontal [PITH_FULL_IMAGE:figures/full_fig_p009_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Visualization of the training evolution of the two specialized branches. [PITH_FULL_IMAGE:figures/full_fig_p010_6.png] view at source ↗
Figure 8
Figure 8. Figure 8: Qualitative comparison of different branch coordination strategies [PITH_FULL_IMAGE:figures/full_fig_p011_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Qualitative comparison of different supervision strategies in the [PITH_FULL_IMAGE:figures/full_fig_p012_9.png] view at source ↗
Figure 11
Figure 11. Figure 11: Underwater robots used for practical application of DSS-USOD. [PITH_FULL_IMAGE:figures/full_fig_p012_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Underwater robotic visual target inspection. (a) Experimental setup. [PITH_FULL_IMAGE:figures/full_fig_p013_12.png] view at source ↗
read the original abstract

Underwater salient object detection (USOD) has attracted increasing attention for underwater visual scene understanding and vision-guided robotic applications. However, existing USOD methods still struggle with underwater image degradations, which often lead to inaccurate object localization, fragmented salient regions, and coarse boundary prediction. To address these challenges, this paper proposes DSS-USOD, a novel RGB-based USOD method built upon dynamic structural specialization. DSS-USOD extracts a shared base representation from a single underwater image, decomposes it into boundary-sensitive and region-coherent structural features, and dynamically coordinates their contributions according to local structural context. Specifically, the extracted shared base representation is decomposed into a boundary-sensitive branch for modeling fine-grained boundary details and a region-coherent branch for capturing region-level structural consistency. A spatial coordination module is then introduced to adaptively regulate the relative contributions of the two branches according to local structural context. Moreover, cooperative structural supervision is introduced to promote branch specialization and stabilize spatial coordination, enabling DSS-USOD to better balance boundary precision and region coherence under degraded underwater conditions. Extensive experiments show that DSS-USOD achieves superior performance on benchmark datasets. Finally, real-world deployment on an underwater robot validates the practical effectiveness of DSS-USOD for underwater object inspection.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The manuscript presents DSS-USOD, a novel RGB-based method for underwater salient object detection. It extracts a shared base representation from a single underwater image, decomposes it into a boundary-sensitive branch for modeling fine-grained boundary details and a region-coherent branch for capturing region-level structural consistency, and introduces a spatial coordination module to adaptively regulate the relative contributions of the two branches according to local structural context. Cooperative structural supervision is proposed to promote branch specialization and stabilize coordination. The central claims are that this architecture corrects inaccurate localization, fragmented regions, and coarse boundaries caused by underwater degradations, achieves superior performance on benchmark datasets, and demonstrates practical effectiveness via real-world deployment on an underwater robot.

Significance. If the dynamic structural specialization and spatial coordination reliably improve boundary precision and region coherence under degraded conditions, the work could advance USOD for vision-guided robotic applications by offering a targeted architectural solution to common underwater imaging challenges. The real-world robot deployment provides additional practical value beyond benchmark results.

major comments (2)
  1. [Method (spatial coordination module description)] The load-bearing assumption is that the spatial coordination module can correctly estimate local structural context (boundary vs. interior) from the same low-contrast, blurred, and color-distorted features that originally cause localization and boundary errors. The manuscript provides no analysis, visualization, or ablation demonstrating that the module avoids misestimation under these conditions; without such evidence the claimed corrective benefit of the branch decomposition and cooperative supervision remains unverified.
  2. [Abstract and Experiments section] The abstract asserts superior performance on benchmark datasets and practical effectiveness via robot deployment, yet the provided text contains no quantitative metrics, baseline comparisons, ablation results on the branches or coordination module, or error analysis. These details are required to substantiate the central performance claims.
minor comments (2)
  1. [Abstract] The abstract is clear but would be strengthened by including one or two key quantitative results (e.g., mIoU or F-measure gains) to convey the magnitude of improvement immediately.
  2. [Notation and figures] Terminology such as 'boundary-sensitive branch' and 'region-coherent branch' should be used consistently in all figures and equations to prevent minor ambiguity.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We sincerely thank the referee for the constructive feedback and the recommendation for major revision. We have carefully reviewed the comments on the spatial coordination module and the substantiation of performance claims in the abstract and experiments. We address each point below, indicating where revisions will be incorporated to strengthen the manuscript.

read point-by-point responses

  1. Referee: [Method (spatial coordination module description)] The load-bearing assumption is that the spatial coordination module can correctly estimate local structural context (boundary vs. interior) from the same low-contrast, blurred, and color-distorted features that originally cause localization and boundary errors. The manuscript provides no analysis, visualization, or ablation demonstrating that the module avoids misestimation under these conditions; without such evidence the claimed corrective benefit of the branch decomposition and cooperative supervision remains unverified.

    Authors: We agree that direct evidence for the spatial coordination module's robustness to underwater degradations is important for validating the overall approach. The current manuscript includes overall architecture ablations and qualitative results, but lacks targeted visualizations of the estimated local structural context or isolated ablations of the module under low-contrast and blurred conditions. In the revised manuscript, we will add visualizations of the coordination maps on representative degraded images and include a dedicated ablation evaluating the module's impact on boundary and region metrics in challenging subsets of the data. This will help confirm that the module contributes to the claimed corrective benefits without misestimation. revision: yes

  2. Referee: [Abstract and Experiments section] The abstract asserts superior performance on benchmark datasets and practical effectiveness via robot deployment, yet the provided text contains no quantitative metrics, baseline comparisons, ablation results on the branches or coordination module, or error analysis. These details are required to substantiate the central performance claims.

    Authors: The abstract serves as a high-level summary and conventionally omits specific numerical results. The full manuscript reports quantitative comparisons against state-of-the-art methods on benchmark datasets (Tables 1–2), ablation studies on the boundary-sensitive branch, region-coherent branch, and spatial coordination module (Section 4.3), as well as qualitative error analysis via visual examples (Figures 3–5). The robot deployment results are presented in Section 5. To better align with the referee's request, we will revise the abstract to include brief mentions of key performance gains (e.g., improvements in mIoU and boundary F-measure) and strengthen cross-references to the experimental sections. We will also expand the error analysis subsection if space allows. revision: partial

Circularity Check

0 steps flagged

No circularity: architectural proposal is self-contained

full rationale

The paper describes DSS-USOD as extracting a shared base representation from an underwater image, decomposing it into boundary-sensitive and region-coherent branches, then using a spatial coordination module and cooperative structural supervision to adaptively balance them. No equations, fitted parameters, or derivations are shown that reduce any claimed prediction or result to quantities defined by the inputs themselves. No self-citations are invoked as load-bearing uniqueness theorems or ansatzes. The central claims rest on empirical benchmark results and robot deployment rather than tautological reductions, making the derivation chain independent and non-circular.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 3 invented entities

The central claim rests on the unverified effectiveness of the newly introduced boundary-sensitive branch, region-coherent branch, spatial coordination module, and cooperative supervision signals for counteracting underwater image degradations; no independent evidence or external benchmarks for these components are supplied in the abstract.

axioms (1)
  • domain assumption Underwater images suffer from degradations that cause inaccurate object localization, fragmented salient regions, and coarse boundary prediction.
    Explicitly stated as the motivation and challenge the method is designed to solve.
invented entities (3)
  • boundary-sensitive branch no independent evidence
    purpose: modeling fine-grained boundary details
    New architectural component introduced to address boundary prediction issues.
  • region-coherent branch no independent evidence
    purpose: capturing region-level structural consistency
    New architectural component introduced to address region fragmentation.
  • spatial coordination module no independent evidence
    purpose: adaptively regulating relative contributions of the two branches according to local structural context
    New module introduced to dynamically balance the branches.

pith-pipeline@v0.9.0 · 5770 in / 1411 out tokens · 44425 ms · 2026-05-19T14:20:50.302742+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

106 extracted references · 106 canonical work pages · 1 internal anchor

  1. [1]

    Perceptual inference, learning, and attention in a multi- sensory world,

    U. Noppeney, “Perceptual inference, learning, and attention in a multi- sensory world,”Annual Review of Neuroscience, vol. 44, pp. 449–473, 2021

    work page 2021

  2. [2]

    A model of saliency-based visual attention for rapid scene analysis,

    L. Itti, C. Koch, and E. Niebur, “A model of saliency-based visual attention for rapid scene analysis,”IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 20, no. 11, pp. 1254–1259, 1998

    work page 1998

  3. [3]

    Global contrast based salient region detection,

    M.-M. Cheng, N. J. Mitra, X. Huang, P. H. Torr, and S.-M. Hu, “Global contrast based salient region detection,”IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 37, no. 3, pp. 569–582, 2014

    work page 2014

  4. [4]

    Deep learning-based marine big data fusion for ocean environment monitoring: Towards shape optimization and salient objects detection,

    S. Khan, I. Ullah, F. Ali, M. Shafiq, Y . Y . Ghadi, and T. Kim, “Deep learning-based marine big data fusion for ocean environment monitoring: Towards shape optimization and salient objects detection,” Frontiers in Marine Science, vol. 9, p. 1094915, 2023

    work page 2023

  5. [5]

    Saliency ranking for benthic survey using underwater images,

    M. Johnson-Roberson, O. Pizarro, and S. Williams, “Saliency ranking for benthic survey using underwater images,” in2010 11th Interna- tional Conference on Control Automation Robotics & Vision. IEEE, 2010, pp. 459–466

    work page 2010

  6. [6]

    Vision- based underwater inspection with portable autonomous underwater vehicle: Development, control, and evaluation,

    L. Hong, X. Wang, D.-S. Zhang, M. Zhao, and H. Xu, “Vision- based underwater inspection with portable autonomous underwater vehicle: Development, control, and evaluation,”IEEE Transactions on Intelligent Vehicles, vol. 9, no. 1, pp. 2197–2209, 2024

    work page 2024

  7. [7]

    Robust hybrid visual servoing for hovering control of autonomous underwater vehicles in unstructured environments,

    L. Hong, X. Wang, and D. Zhang, “Robust hybrid visual servoing for hovering control of autonomous underwater vehicles in unstructured environments,”Ocean Engineering, vol. 339, p. 122103, 2025

    work page 2025

  8. [8]

    Sea-thru: A method for removing water from underwater images,

    D. Akkaynak and T. Treibitz, “Sea-thru: A method for removing water from underwater images,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2019, pp. 1682–1691

    work page 2019

  9. [9]

    Wsuie: Weakly supervised underwater image enhancement for improved visual per- ception,

    L. Hong, X. Wang, Z. Xiao, G. Zhang, and J. Liu, “Wsuie: Weakly supervised underwater image enhancement for improved visual per- ception,”IEEE Robotics and Automation Letters, vol. 6, no. 4, pp. 8237–8244, 2021

    work page 2021

  10. [10]

    Underwater salient object detection via dual-stage self-paced learning and depth emphasis,

    J. Jin, Q. Jiang, Q. Wu, B. Xu, and R. Cong, “Underwater salient object detection via dual-stage self-paced learning and depth emphasis,”IEEE Transactions on Circuits and Systems for Video Technology, 2024

    work page 2024

  11. [11]

    Gradient-based learning applied to document recognition,

    Y . LeCun, L. Bottou, Y . Bengio, and P. Haffner, “Gradient-based learning applied to document recognition,”Proceedings of the IEEE, vol. 86, no. 11, pp. 2278–2324, 1998

    work page 1998

  12. [12]

    An image is worth 16x16 words: Trans- formers 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, J. Uszkoreit, and N. Houlsby, “An image is worth 16x16 words: Trans- formers for image recognition at scale,” inInternational Conference on Learning Representations, 2021

    work page 2021

  13. [13]

    Salient object detection: A benchmark,

    A. Borji, M.-M. Cheng, H. Jiang, and J. Li, “Salient object detection: A benchmark,”IEEE Transactions on Image Processing, vol. 24, no. 12, pp. 5706–5722, 2015

    work page 2015

  14. [14]

    Rgb-d salient object detection: A survey,

    T. Zhou, D.-P. Fan, M.-M. Cheng, J. Shen, and L. Shao, “Rgb-d salient object detection: A survey,”Computational Visual Media, pp. 1–33, 2021

    work page 2021

  15. [15]

    Effiseanet: Pio- neering lightweight network for underwater salient object detection,

    Q. Wu, Z. Fu, H. Lin, C. Ma, X. Tu, and X. Ding, “Effiseanet: Pio- neering lightweight network for underwater salient object detection,” inProceedings of the Asian Conference on Computer Vision, 2024, pp. 1486–1501

    work page 2024

  16. [16]

    A fusion underwater salient object detection based on multi-scale saliency and spatial optimization,

    W. Huang, X. Zhuet al., “A fusion underwater salient object detection based on multi-scale saliency and spatial optimization,”Journal of Marine Science and Engineering, vol. 11, no. 9, p. 1757, 2023

    work page 2023

  17. [17]

    Ce 3usod: Channel-enhanced, efficient, and effective network for underwater salient object detection,

    Q. Wu, J. Xie, Z. Fu, X. Tu, Y . Huang, and X. Ding, “Ce 3usod: Channel-enhanced, efficient, and effective network for underwater salient object detection,”IEEE Journal of Oceanic Engineering, vol. 50, no. 2, pp. 941–954, 2025

    work page 2025

  18. [18]

    If-usod: Multimodal information fusion interactive feature enhancement architecture for underwater salient object detection,

    G. Yuan, J. Song, and J. Li, “If-usod: Multimodal information fusion interactive feature enhancement architecture for underwater salient object detection,”Information Fusion, vol. 117, p. 102806, 2025

    work page 2025

  19. [19]

    Detecting underwater salient objects via self-supervised depth priors and task-driven optimization,

    Y . Liu, X. Zhang, K. Zhang, B. Ma, S. Yang, R. Yang, and P. Tan, “Detecting underwater salient objects via self-supervised depth priors and task-driven optimization,”Expert Systems with Applications, p. 130873, 2025

    work page 2025

  20. [20]

    Udepth: Fast monocular depth estima- tion for visually-guided underwater robots,

    B. Yu, J. Wu, and M. J. Islam, “Udepth: Fast monocular depth estima- tion for visually-guided underwater robots,” in2023 IEEE International Conference on Robotics and Automation. IEEE, 2023, pp. 3116–3123

    work page 2023

  21. [21]

    Usod10k: A new benchmark dataset for underwater salient object detection,

    L. Hong, X. Wang, G. Zhang, and M. Zhao, “Usod10k: A new benchmark dataset for underwater salient object detection,”IEEE Transactions on Image Processing, vol. 34, pp. 1602–1615, 2025

    work page 2025

  22. [22]

    Calibrated rgb-d salient object detection,

    W. Ji, J. Li, S. Yu, M. Zhang, Y . Piao, S. Yao, Q. Bi, K. Ma, Y . Zheng, H. Lu, and L. Cheng, “Calibrated rgb-d salient object detection,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, June 2021, pp. 9471–9481

    work page 2021

  23. [23]

    Accurate rgb-d salient object detection via collaborative learning,

    W. Ji, J. Li, M. Zhang, Y . Piao, and H. Lu, “Accurate rgb-d salient object detection via collaborative learning,” inProceedings of the European Conference on Computer Vision, 2020, pp. 52–69

    work page 2020

  24. [24]

    Basnet: Boundary-aware salient object detection,

    X. Qin, Z. Zhang, C. Huang, C. Gao, M. Dehghan, and M. Jagersand, “Basnet: Boundary-aware salient object detection,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recogni- tion, 2019, pp. 7479–7489

    work page 2019

  25. [25]

    Edge-guided non-local fully convolutional network for salient object detection,

    Z. Tu, Y . Ma, C. Li, J. Tang, and B. Luo, “Edge-guided non-local fully convolutional network for salient object detection,”IEEE Transactions on Circuits and Systems for Video Technology, vol. 31, no. 2, pp. 582– 593, 2020

    work page 2020

  26. [26]

    Egnet: Edge guidance network for salient object detection,

    J. Zhao, J.-J. Liu, D.-P. Fan, Y . Cao, J. Yang, and M.-M. Cheng, “Egnet: Edge guidance network for salient object detection,” in2019 IEEE/CVF International Conference on Computer Vision, 2019, pp. 8778–8787

    work page 2019

  27. [27]

    Selectivity or invari- ance: Boundary-aware salient object detection,

    J. Su, J. Li, Y . Zhang, C. Xia, and Y . Tian, “Selectivity or invari- ance: Boundary-aware salient object detection,” inProceedings of the IEEE/CVF International Conference on Computer Vision, 2019, pp. 3799–3808

    work page 2019

  28. [28]

    Label de- coupling framework for salient object detection,

    J. Wei, S. Wang, Z. Wu, C. Su, Q. Huang, and Q. Tian, “Label de- coupling framework for salient object detection,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2020, pp. 13 025–13 034. IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE 14

    work page 2020

  29. [29]

    Csunet: Contour-sensitive underwater salient object detection,

    Y . Wei, Y . Wang, S. Yan, T. Wang, Z. Wang, W. Sun, Y . Zhao, and X. Xue, “Csunet: Contour-sensitive underwater salient object detection,” inProceedings of the 6th ACM Multimedia Asia, 2024, pp. 78:1–78:7

    work page 2024

  30. [30]

    Edge distraction-aware salient object detection,

    S. Ren, W. Liu, J. Jiao, G. Han, and S. He, “Edge distraction-aware salient object detection,”IEEE MultiMedia, vol. 30, no. 3, pp. 63–73, 2023

    work page 2023

  31. [31]

    Filling-in the forms: Surface and boundary interactions in visual cortex,

    S. Grossberg, “Filling-in the forms: Surface and boundary interactions in visual cortex,” inFilling-in: From Perceptual Completion to Skill Learning, L. Pessoa and P. D. Weerd, Eds. New York: Oxford University Press, 2003, pp. 13–37

    work page 2003

  32. [32]

    Mechanisms of visual attention in the human cortex,

    S. Kastner and L. G. Ungerleider, “Mechanisms of visual attention in the human cortex,”Annual Review of Neuroscience, vol. 23, pp. 315– 341, 2000

    work page 2000

  33. [33]

    Under- water salient object detection by combining 2d and 3d visual features,

    Z. Chen, H. Gao, Z. Zhang, H. Zhou, X. Wang, and Y . Tian, “Under- water salient object detection by combining 2d and 3d visual features,” Neurocomputing, vol. 391, pp. 249–259, 2020

    work page 2020

  34. [34]

    Salient object detection in the deep learning era: An in-depth survey,

    W. Wang, Q. Lai, H. Fu, J. Shen, H. Ling, and R. Yang, “Salient object detection in the deep learning era: An in-depth survey,”IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 44, no. 6, pp. 3239–3259, 2021

    work page 2021

  35. [35]

    Hdanet: Enhancing underwater salient object detection with physics-inspired multimodal joint learning,

    Y . Liu, X. Zhang, J. Zhu, B. Ma, Y . Duan, and P. Tan, “Hdanet: Enhancing underwater salient object detection with physics-inspired multimodal joint learning,”IEEE Transactions on Geoscience and Remote Sensing, 2025

    work page 2025

  36. [36]

    Turbidity–similarity decoupling: Feature-consistent mutual learning for underwater salient object detection,

    W. Zhou, B. Tang, R. Cong, and Q. Jiang, “Turbidity–similarity decoupling: Feature-consistent mutual learning for underwater salient object detection,”IEEE Transactions on Image Processing, pp. 1–1, 2026

    work page 2026

  37. [37]

    Blurriness-guided underwater salient object detection and data augmentation,

    Y .-T. Peng, Y .-C. Lin, W.-Y . Peng, and C.-Y . Liu, “Blurriness-guided underwater salient object detection and data augmentation,”IEEE Journal of Oceanic Engineering, vol. 49, no. 3, pp. 1089–1103, 2024

    work page 2024

  38. [38]

    Heterogeneous experts and hierarchical perception for underwater salient object detection,

    M. Zha, G. Wang, Y . Pei, T. Li, X. Tang, C. Li, Y . Yang, and H. T. Shen, “Heterogeneous experts and hierarchical perception for underwater salient object detection,”IEEE Transactions on Image Processing, 2025

    work page 2025

  39. [39]

    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,” inPro- ceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2019, pp. 3917–3926

    work page 2019

  40. [40]

    Multi-scale interactive network for salient object detection,

    Y . Pang, X. Zhao, L. Zhang, and H. Lu, “Multi-scale interactive network for salient object detection,” in2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2020, pp. 9410–9419

    work page 2020

  41. [41]

    F 3net: fusion, feedback and focus for salient object detection,

    J. Wei, S. Wang, and Q. Huang, “F 3net: fusion, feedback and focus for salient object detection,” inProceedings of the AAAI conference on artificial intelligence, vol. 34, no. 07, 2020, pp. 12 321–12 328

    work page 2020

  42. [42]

    Stacked cross refinement network for edge-aware salient object detection,

    Z. Wu, L. Su, and Q. Huang, “Stacked cross refinement network for edge-aware salient object detection,” inProceedings of the IEEE/CVF International Conference on Computer Vision, 2019, pp. 7264–7273

    work page 2019

  43. [43]

    Pytorch: An imperative style, high-performance deep learning library,

    A. Paszke, S. Gross, F. Massa, A. Lerer, J. Bradbury, G. Chanan, T. Killeen, Z. Lin, N. Gimelshein, L. Antigaet al., “Pytorch: An imperative style, high-performance deep learning library,”Advances in neural information processing systems, vol. 32, pp. 8026–8037, 2019

    work page 2019

  44. [44]

    Pvt v2: Improved baselines with pyramid vision transformer,

    W. Wang, E. Xie, X. Li, D.-P. Fan, K. Song, D. Liang, T. Lu, P. Luo, and L. Shao, “Pvt v2: Improved baselines with pyramid vision transformer,”Computational visual media, vol. 8, no. 3, pp. 415–424, 2022

    work page 2022

  45. [45]

    Adam: A Method for Stochastic Optimization

    D. P. Kingma and J. Ba, “Adam: A method for stochastic optimization,” arXiv preprint arXiv:1412.6980, 2014

    work page internal anchor Pith review Pith/arXiv arXiv 2014

  46. [46]

    Svam: Saliency-guided visual at- tention modeling by autonomous underwater robots,

    M. J. Islam, R. Wang, and J. Sattar, “Svam: Saliency-guided visual at- tention modeling by autonomous underwater robots,” in18th Robotics: Science and Systems, RSS 2022. MIT Press Journals, 2022

    work page 2022

  47. [47]

    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,” in2017 IEEE International Conference on Computer Vision, 2017, pp. 4558–4567

    work page 2017

  48. [48]

    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,” inProceedings of the 27th International Joint Conference on Artificial Intelligence, 2018, pp. 698–704

    work page 2018

  49. [49]

    Frequency-tuned salient region detection,

    R. Achanta, S. Hemami, F. Estrada, and S. Susstrunk, “Frequency-tuned salient region detection,” in2009 IEEE Conference on Computer Vision and Pattern Recognition, 2009, pp. 1597–1604

    work page 2009

  50. [50]

    Saliency filters: Contrast based filtering for salient region detection,

    F. Perazzi, P. Kr ¨ahenb¨uhl, Y . Pritch, and A. Hornung, “Saliency filters: Contrast based filtering for salient region detection,” in2012 IEEE Conference on Computer Vision and Pattern Recognition, 2012, pp. 733–740

    work page 2012

  51. [51]

    Progressive feature polishing network for salient object detection,

    B. Wang, Q. Chen, M. Zhou, Z. Zhang, and K. Gai, “Progressive feature polishing network for salient object detection,”Proceedings of the AAAI Conference on Artificial Intelligence, vol. 34, no. 7, pp. 12 128–12 135, 2020

    work page 2020

  52. [52]

    Is depth really necessary for salient object detection?

    J. Zhao, Y . Zhao, J. Li, and X. Chen, “Is depth really necessary for salient object detection?” inProceedings of the 28th ACM International Conference on Multimedia, 2020, pp. 1745–1754

    work page 2020

  53. [53]

    Pyramidal feature shrinking for salient object detection,

    M. Ma, C. Xia, and J. Li, “Pyramidal feature shrinking for salient object detection,” inProceedings of the AAAI Conference on Artificial Intelligence, vol. 35, no. 3, 2021, pp. 2311–2318

    work page 2021

  54. [54]

    Mfnet: Multi-filter directive network for weakly supervised salient object detection,

    Y . Piao, J. Wang, M. Zhang, and H. Lu, “Mfnet: Multi-filter directive network for weakly supervised salient object detection,” inProceedings of the IEEE/CVF International Conference on Computer Vision, 2021, pp. 4136–4145

    work page 2021

  55. [55]

    Complementary trilateral decoder for fast and accurate salient object detection,

    Z. Zhao, C. Xia, C. Xie, and J. Li, “Complementary trilateral decoder for fast and accurate salient object detection,” inProceedings of the 29th ACM International Conference on Multimedia, 2021, pp. 4967– 4975

    work page 2021

  56. [56]

    Progressive self- guided loss for salient object detection,

    S. Yang, W. Lin, G. Lin, Q. Jiang, and Z. Liu, “Progressive self- guided loss for salient object detection,”IEEE Transactions on Image Processing, vol. 30, pp. 8426–8438, 2021

    work page 2021

  57. [57]

    Visual saliency trans- former,

    N. Liu, N. Zhang, K. Wan, L. Shao, and J. Han, “Visual saliency trans- former,” inProceedings of the IEEE/CVF International Conference on Computer Vision, October 2021, pp. 4722–4732

    work page 2021

  58. [58]

    A highly efficient model to study the semantics of salient object detec- tion,

    M.-M. Cheng, S.-H. Gao, A. Borji, Y .-Q. Tan, Z. Lin, and M. Wang, “A highly efficient model to study the semantics of salient object detec- tion,”IEEE Transactions on Pattern Analysis and Machine Intelligence, 2022

    work page 2022

  59. [59]

    Separate first, then segment: An integrity segmentation network for salient object detection,

    G. Zhu, J. Li, and Y . Guo, “Separate first, then segment: An integrity segmentation network for salient object detection,”Pattern Recognition, vol. 150, p. 110328, 2024

    work page 2024

  60. [60]

    Boosting salient object detection with transformer-based asymmetric bilateral u-net,

    Y . Qiu, Y . Liu, L. Zhang, H. Lu, and J. Xu, “Boosting salient object detection with transformer-based asymmetric bilateral u-net,”IEEE Transactions on Circuits and Systems for Video Technology, vol. 34, no. 4, 2024

    work page 2024

  61. [61]

    Genera- tive transformer for accurate and reliable salient object detection,

    Y . Mao, J. Zhang, Z. Wan, X. Tian, A. Li, Y . Lv, and Y . Dai, “Genera- tive transformer for accurate and reliable salient object detection,”IEEE Transactions on Circuits and Systems for Video Technology, vol. 35, no. 2, pp. 1041–1054, 2025

    work page 2025

  62. [62]

    Rapid salient object detection with difference con- volutional neural networks,

    Z. Su, L. Liu, M. M ¨uller, J. Zhang, D. Wofk, M.-M. Cheng, and M. Pietik ¨ainen, “Rapid salient object detection with difference con- volutional neural networks,”IEEE Transactions on Pattern Analysis and Machine Intelligence, 2025

    work page 2025

  63. [63]

    Jl-dcf: Joint learning and densely-cooperative fusion framework for rgb-d salient object detection,

    K. Fu, D.-P. Fan, G.-P. Ji, and Q. Zhao, “Jl-dcf: Joint learning and densely-cooperative fusion framework for rgb-d salient object detection,” inProceedings of the IEEE/CVF conference on Computer Vision and Pattern Recognition, 2020, pp. 3052–3062

    work page 2020

  64. [64]

    Uc-net: Uncertainty inspired rgb-d saliency detection via conditional variational autoencoders,

    J. Zhang, D.-P. Fan, Y . Dai, S. Anwar, F. Sadat Saleh, T. Zhang, and N. Barnes, “Uc-net: Uncertainty inspired rgb-d saliency detection via conditional variational autoencoders,” inProceedings of the IEEE conference on Computer Vision and Pattern Recognition, 2020

    work page 2020

  65. [65]

    Learning selective mutual attention and contrast for rgb-d saliency detection,

    N. Liu, N. Zhang, L. Shao, and J. Han, “Learning selective mutual attention and contrast for rgb-d saliency detection,”IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 44, no. 12, pp. 9026–9042, 2021

    work page 2021

  66. [66]

    Bbs-net: Rgb-d salient object detection with a bifurcated backbone strategy network,

    D.-P. Fan, Y . Zhai, A. Borji, J. Yang, and L. Shao, “Bbs-net: Rgb-d salient object detection with a bifurcated backbone strategy network,” inEuropean Conference on Computer Vision. Springer, 2020, pp. 275–292

    work page 2020

  67. [67]

    A single stream network for robust and real-time rgb-d salient object detection,

    X. Zhao, L. Zhang, Y . Pang, H. Lu, and L. Zhang, “A single stream network for robust and real-time rgb-d salient object detection,” in European Conference on Computer Vision. Springer, 2020, pp. 646– 662

    work page 2020

  68. [68]

    Specificity- preserving rgb-d saliency detection,

    T. Zhou, H. Fu, G. Chen, Y . Zhou, D.-P. Fan, and L. Shao, “Specificity- preserving rgb-d saliency detection,” inProceedings of the IEEE/CVF International Conference on Computer Vision, 2021, pp. 4681–4691

    work page 2021

  69. [69]

    Hierarchical alternate interaction network for rgb-d salient object detection,

    G. Li, Z. Liu, M. Chen, Z. Bai, W. Lin, and H. Ling, “Hierarchical alternate interaction network for rgb-d salient object detection,”IEEE Transactions on Image Processing, vol. 30, pp. 3528–3542, 2021

    work page 2021

  70. [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,” Proceedings of the 29th ACM International Conference on Multimedia, 2021

    work page 2021

  71. [71]

    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. IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE 15

    work page 2075

  72. [72]

    Bts-net: Bi-directional transfer-and-selection network for rgb-d salient object detection,

    W. Zhang, Y . Jiang, K. Fu, and Q. Zhao, “Bts-net: Bi-directional transfer-and-selection network for rgb-d salient object detection,” in 2021 IEEE International Conference on Multimedia and Expo. IEEE, 2021, pp. 1–6

    work page 2021

  73. [73]

    Cross-modality discrepant interaction network for RGB-D salient object detection,

    C. Zhang, R. Cong, Q. Lin, L. Ma, L. Feng, Y . Zhao, and S. Kwong, “Cross-modality discrepant interaction network for RGB-D salient object detection,” inProceedings of the 29th ACM International Conference on Multimedia. ACM, 2021

    work page 2021

  74. [74]

    Cir-net: Cross-modality interaction and refinement for rgb-d salient object detection,

    R. Cong, Q. Lin, C. Zhang, C. Li, X. Cao, Q. Huang, and Y . Zhao, “Cir-net: Cross-modality interaction and refinement for rgb-d salient object detection,”IEEE Transactions on Image Processing, vol. 31, 2022

    work page 2022

  75. [75]

    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, 2023

    work page 2023

  76. [76]

    Point-aware interaction and cnn-induced refinement net- work 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 net- work for rgb-d salient object detection,” inProceedings of the 31st ACM International Conference on Multimedia, 2023

    work page 2023

  77. [77]

    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, vol. 26, 2024

    work page 2024

  78. [78]

    Lightweight rgb-d salient object detection from a speed-accuracy tradeoff perspective,

    S. Duan, X. Yang, N. Wang, and X. Gao, “Lightweight rgb-d salient object detection from a speed-accuracy tradeoff perspective,”IEEE Transactions on Image Processing, vol. 34, pp. 2529–2543, 2025

    work page 2025

  79. [79]

    Fscdiff: Frequency- spatial entangled conditional diffusion model for underwater salient object detection,

    H. Li, G. Lin, Z. Li, S. Kwong, and R. Cong, “Fscdiff: Frequency- spatial entangled conditional diffusion model for underwater salient object detection,” inProceedings of the 33rd ACM International Conference on Multimedia, 2025, pp. 8379–8388

    work page 2025

  80. [80]

    Deep residual learning for image recognition,

    K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for image recognition,” inProceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2016, pp. 770–778

    work page 2016

Showing first 80 references.