pith. sign in

arxiv: 2607.01825 · v1 · pith:GIYMFKBTnew · submitted 2026-07-02 · 💻 cs.CV

Rethinking Conditional Generation for Underwater Salient Object Detection

Hua Li , Yongjie Weng , Yutong Li , Zhiyuan Li , Runmin Cong , Sam Kwong This is my paper

Pith reviewed 2026-07-03 16:06 UTC · model grok-4.3

classification 💻 cs.CV
keywords underwater salient object detectionconditional generationphysics priordegradation-aware featuresdiffusion transformerpseudo-depth guidance
0
0 comments X

The pith

DCGNet restores degradation-aware RGB features from pseudo-depth to improve underwater salient object detection.

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

The paper presents a Degradation-aware Conditional Generation Network (DCGNet) that builds reliable conditional features for generating saliency maps in underwater images. It combines a Dynamic Multi-Granularity module for scale-varying objects, an Underwater Physics-Prior module that estimates light attenuation and backscatter via pseudo-depth, a spatial Gaussian saliency prior, and a diffusion transformer bottleneck. A sympathetic reader would care because scattering and absorption effects break conventional saliency methods, so a physics-informed conditional approach could make detection viable in marine settings. Experiments on five datasets show consistent gains over prior methods.

Core claim

DCGNet constructs reliable conditional features for underwater saliency generation through an Underwater Physics-Prior module that uses pseudo-depth guidance to estimate light attenuation and backscatter, restoring degradation-aware RGB features, together with a Dynamic Multi-Granularity module for blurred boundaries, an Underwater Spatial Gaussian module for object-centered priors, and a timestep-adaptive Diffusion Transformer in the denoising decoder; this yields superior performance on USOD10K, USOD, CSOD10K, MAS3K, and RMAS.

What carries the argument

The Underwater Physics-Prior module (UPP), which estimates underwater light attenuation and backscatter from pseudo-depth guidance to restore degradation-aware RGB features.

If this is right

  • DCGNet outperforms existing state-of-the-art methods on USOD10K, USOD, CSOD10K, MAS3K, and RMAS.
  • The Dynamic Multi-Granularity module detects salient objects of varying scales despite blurred boundaries.
  • The Underwater Spatial Gaussian module enhances object-centered regions while suppressing cluttered backgrounds.
  • The timestep-adaptive Diffusion Transformer refines fused features across diffusion steps.
  • The overall architecture shows potential for complex underwater visual applications.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Similar pseudo-depth-guided restoration could be tested on other scattering-heavy domains such as foggy or hazy scenes.
  • The conditional generation framing suggests that diffusion-based decoders may benefit from domain physics at multiple timesteps.
  • If pseudo-depth estimation itself improves, the downstream saliency gains would likely increase without changing the rest of the network.

Load-bearing premise

Pseudo-depth guidance supplies enough information to estimate light attenuation and backscatter accurately enough to restore useful RGB features.

What would settle it

On a new underwater dataset with ground-truth saliency maps, if DCGNet does not produce higher accuracy metrics than the previous best method, the claim that the physics-prior restoration enables better detection would be undermined.

Figures

Figures reproduced from arXiv: 2607.01825 by Hua Li, Runmin Cong, Sam Kwong, Yongjie Weng, Yutong Li, Zhiyuan Li.

Figure 1
Figure 1. Figure 1: Visualization of conditional feature reliability in USOD. The first row [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Overall architecture of DCGNet. The Condition Net extracts multi-level features through PVTv2 encoder stages and repeated DCG Blocks (Degradation [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Overall architecture of Dynamic Multi-Granularity. By dynamically generating large-kernel weights, performing multi-scale feature aggregation, and [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Overview of the Underwater Spatial Gaussian (USG) module. Given [PITH_FULL_IMAGE:figures/full_fig_p005_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Qualitative comparison on the USOD10K [7] and USOD [14] benchmarks. The examples cover low contrast, severe color degradation, weak foreground [PITH_FULL_IMAGE:figures/full_fig_p008_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Qualitative comparison on the CSOD10K benchmark. The examples [PITH_FULL_IMAGE:figures/full_fig_p009_6.png] view at source ↗
Figure 8
Figure 8. Figure 8: Qualitative ablation results on the USOD10K benchmark, showing [PITH_FULL_IMAGE:figures/full_fig_p010_8.png] view at source ↗
read the original abstract

Salient Object Detection in underwater images remains challenging due to low contrast, uneven illumination, and color distortion caused by scattering and absorption effects, which limit the effectiveness of conventional SOD methods in underwater environments. To address these challenges, we propose a Degradation-aware Conditional Generation Network (DCGNet), specifically designed to construct reliable conditional features for underwater saliency generation. First, we design a Dynamic Multi-Granularity module (DMG) grounded in the human visual system to robustly detect salient objects of varying scales with blurred boundaries. Then, we develop an Underwater Physics-Prior module (UPP), which utilizes pseudo-depth guidance to estimate underwater light attenuation and backscatter, thereby restoring degradation-aware RGB features and mitigating color distortion and boundary ambiguity. Based on the physics-guided representation, we introduce an Underwater Spatial Gaussian module (USG), which constructs a spatial Gaussian saliency prior from the strongest guided response to enhance object-centered salient regions and suppress cluttered underwater backgrounds. In addition, a lightweight timestep-adaptive Diffusion Transformer (DiT) bottleneck is inserted into the denoising decoder to refine fused features at different diffusion timesteps. Comprehensive experiments on USOD10K, USOD, CSOD10K, MAS3K, and RMAS demonstrate that DCGNet significantly outperforms existing state-of-the-art methods, verifying its potential for complex underwater visual applications.

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

1 major / 2 minor

Summary. The manuscript proposes DCGNet, a Degradation-aware Conditional Generation Network for underwater salient object detection. It introduces four components: a Dynamic Multi-Granularity (DMG) module for detecting salient objects at varying scales with blurred boundaries, an Underwater Physics-Prior (UPP) module that uses pseudo-depth guidance to estimate light attenuation and backscatter for restoring degradation-aware RGB features, an Underwater Spatial Gaussian (USG) module to build a spatial Gaussian saliency prior from the strongest guided response, and a lightweight timestep-adaptive Diffusion Transformer (DiT) bottleneck in the denoising decoder. The central claim is that this architecture significantly outperforms existing state-of-the-art methods on the USOD10K, USOD, CSOD10K, MAS3K, and RMAS datasets.

Significance. If the reported gains prove robust under standard evaluation protocols with ablations and statistical controls, the work could contribute a physics-informed conditional generation pipeline tailored to underwater domain challenges such as scattering and absorption. The explicit use of pseudo-depth for attenuation/backscatter estimation and the insertion of a timestep-adaptive DiT are architecture-specific choices that differentiate it from generic SOD or diffusion baselines.

major comments (1)
  1. [Abstract] Abstract: the assertion that DCGNet 'significantly outperforms existing state-of-the-art methods' on five named datasets is presented without any quantitative metrics, tables, error bars, ablation results, or statistical tests. This absence makes the central empirical claim impossible to evaluate from the supplied text.
minor comments (2)
  1. [Abstract] Abstract: the description of the UPP module states that it 'utilizes pseudo-depth guidance to estimate underwater light attenuation and backscatter' but provides no equation, loss term, or implementation detail for how the estimation is performed or how the restored features are fused.
  2. [Abstract] Abstract: the DiT is described as 'lightweight' and 'timestep-adaptive' without specifying the number of layers, attention mechanism, or how timestep conditioning is injected.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for their comments on our manuscript. We address the single major comment below.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the assertion that DCGNet 'significantly outperforms existing state-of-the-art methods' on five named datasets is presented without any quantitative metrics, tables, error bars, ablation results, or statistical tests. This absence makes the central empirical claim impossible to evaluate from the supplied text.

    Authors: We agree that the abstract, as a concise summary, states the performance claim without accompanying numbers. The full manuscript contains the supporting evidence: Tables 1–5 report quantitative results (mIoU, F-measure, MAE, etc.) on USOD10K, USOD, CSOD10K, MAS3K, and RMAS, with comparisons to prior SOTA methods, plus ablation studies and statistical controls in Section 4. To make the abstract self-contained and address the concern directly, we will revise it in the next version to include the key average gains (e.g., +X% mIoU across datasets) while preserving its brevity. revision: yes

Circularity Check

0 steps flagged

No significant circularity detected

full rationale

The provided abstract and module descriptions introduce architectural components (DMG, UPP, USG, DiT) for conditional generation in underwater SOD. No equations, fitted parameters, predictions, self-citations, uniqueness theorems, or ansatzes are present that would reduce any claimed result to its inputs by construction. The central claim of outperformance is presented as an empirical result on external datasets and is therefore externally falsifiable rather than internally forced. This is the most common honest non-finding for papers whose load-bearing steps are architectural and experimental.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only; no explicit free parameters, axioms, or invented entities are stated. The modules themselves are presented as novel engineering contributions rather than new physical entities.

pith-pipeline@v0.9.1-grok · 5779 in / 1094 out tokens · 24867 ms · 2026-07-03T16:06:09.504467+00:00 · methodology

discussion (0)

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

Reference graph

Works this paper leans on

68 extracted references · 68 canonical work pages

  1. [1]

    Vit-uwa: Vision transformer underwater-adapter for dense predictions beneath the water surface,

    Y . Jia, Q. Lin, H. Li, Y . Li, S. Kwong, and R. Cong, “Vit-uwa: Vision transformer underwater-adapter for dense predictions beneath the water surface,”IEEE Transactions on Image Processing, 2026

    work page 2026

  2. [2]

    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, vol. 35, no. 3, pp. 2147–2160, 2025

    work page 2025

  3. [3]

    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

  4. [4]

    Real-world underwater enhancement: Challenges, benchmarks, and solutions under natural light,

    R. Liu, X. Fan, M. Zhu, M. Hou, and Z. Luo, “Real-world underwater enhancement: Challenges, benchmarks, and solutions under natural light,”IEEE Transactions on Circuits and Systems for Video Technology, vol. 30, no. 12, pp. 4861–4875, 2020

    work page 2020

  5. [5]

    Underwater image enhancement via weighted wavelet visual perception fusion,

    W. Zhang, L. Zhou, P. Zhuang, G. Li, X. Pan, W. Zhao, and C. Li, “Underwater image enhancement via weighted wavelet visual perception fusion,”IEEE Transactions on Circuits and Systems for Video Technol- ogy, vol. 34, no. 4, pp. 2469–2483, 2023

    work page 2023

  6. [6]

    Learning heavily- degraded prior for underwater object detection,

    C. Fu, X. Fan, J. Xiao, W. Yuan, R. Liu, and Z. Luo, “Learning heavily- degraded prior for underwater object detection,”IEEE Transactions on Circuits and Systems for Video Technology, vol. 33, no. 11, pp. 6887– 6896, 2023

    work page 2023

  7. [7]

    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, 2023

    work page 2023

  8. [8]

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

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

    work page 2025

  9. [9]

    Generative transformer for accurate and reliable salient object detection,

    Y . Mao, J. Zhang, Z. Wan, X. Tian, A. Li, Y . Lv, and Y . Dai, “Generative 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

  10. [10]

    Speed-up DDPM for real-time underwater image enhancement,

    S. Lu, F. Guan, H. Zhang, and H. Lai, “Speed-up DDPM for real-time underwater image enhancement,”IEEE Transactions on Circuits and Systems for Video Technology, vol. 34, no. 5, pp. 3576–3588, 2024

    work page 2024

  11. [11]

    Waterdiffusion: Learn- ing a prior-involved unrolling diffusion for joint underwater saliency detection and visual restoration,

    L. Chang, Y . Wang, L. Deng, B. Du, and C. Xu, “Waterdiffusion: Learn- ing a prior-involved unrolling diffusion for joint underwater saliency detection and visual restoration,” inProceedings of the AAAI Conference on Artificial Intelligence, vol. 39, no. 2, 2025, pp. 1998–2006

    work page 2025

  12. [12]

    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 Con- ference on Multimedia, 2025, pp. 8379–8388

    work page 2025

  13. [13]

    WaterFlow: Explicit physics-prior rectified flow for underwater saliency mask generation,

    R. Li, S. Lian, H. Li, Y . Li, W. Wu, and S. Kwong, “WaterFlow: Explicit physics-prior rectified flow for underwater saliency mask generation,” in ICASSP, 2026

    work page 2026

  14. [14]

    SV AM: Saliency-guided visual attention modeling by autonomous underwater robots,

    M. J. Islam, R. Wang, and J. Sattar, “SV AM: Saliency-guided visual attention modeling by autonomous underwater robots,” inRobotics: Science and Systems, New York, NY , USA, 2022

    work page 2022

  15. [15]

    Turbidity-similarity decou- pling: Feature-consistent mutual learning for underwater salient object detection,

    W. Zhou, B. Tang, R. Cong, and Q. Jiang, “Turbidity-similarity decou- pling: Feature-consistent mutual learning for underwater salient object detection,”IEEE Transactions on Image Processing, vol. 35, pp. 495– 510, 2026

    work page 2026

  16. [16]

    Conditional text image generation with diffusion models,

    Y . Zhu, Z. Li, T. Wang, M. He, and C. Yao, “Conditional text image generation with diffusion models,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023, pp. 14 235–14 245

    work page 2023

  17. [17]

    Codi: Conditional diffusion distillation for higher-fidelity and faster image generation,

    K. Mei, M. Delbracio, H. Talebi, Z. Tu, V . M. Patel, and P. Milanfar, “Codi: Conditional diffusion distillation for higher-fidelity and faster image generation,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 9048–9058

    work page 2024

  18. [18]

    Residual diffusion deblurring model for single image defocus deblurring,

    H. Feng, H. Zhou, T. Ye, S. Chen, and L. Zhu, “Residual diffusion deblurring model for single image defocus deblurring,” inProceedings of the AAAI Conference on Artificial Intelligence, vol. 39, no. 3, 2025, pp. 2960–2968

    work page 2025

  19. [19]

    Hierarchical integration diffusion model for realistic image deblurring,

    Z. Chen, Y . Zhang, D. Liu, J. Gu, L. Kong, X. Yuanet al., “Hierarchical integration diffusion model for realistic image deblurring,”Advances in Neural Information Processing Systems, vol. 36, pp. 29 114–29 125, 2023

    work page 2023

  20. [20]

    Medsegdiff-v2: Diffusion- based medical image segmentation with transformer,

    J. Wu, W. Ji, H. Fu, M. Xu, Y . Jin, and Y . Xu, “Medsegdiff-v2: Diffusion- based medical image segmentation with transformer,” inProceedings of the AAAI Conference on Artificial Intelligence, vol. 38, no. 6, 2024, pp. 6030–6038

    work page 2024

  21. [21]

    Camodiffusion: Camouflaged object detection via conditional diffusion models,

    Z. Chen, K. Sun, and X. Lin, “Camodiffusion: Camouflaged object detection via conditional diffusion models,” inProceedings of the AAAI Conference on Artificial Intelligence, vol. 38, no. 2, 2024, pp. 1272– 1280

    work page 2024

  22. [22]

    Conditional diffusion models for camouflaged and salient object detection,

    K. Sun, Z. Chen, X. Lin, X. Sun, H. Liu, and R. Ji, “Conditional diffusion models for camouflaged and salient object detection,”IEEE Transactions on Pattern Analysis and Machine Intelligence, 2025

    work page 2025

  23. [23]

    Diffusion models, image super-resolution, and everything: A survey,

    B. B. Moser, A. S. Shanbhag, F. Raue, S. Frolov, S. Palacio, and A. Dengel, “Diffusion models, image super-resolution, and everything: A survey,”IEEE Transactions on Neural Networks and Learning Systems, 2024

    work page 2024

  24. [24]

    Scalable diffusion models with transformers,

    W. Peebles and S. Xie, “Scalable diffusion models with transformers,” inProceedings of the IEEE/CVF International Conference on Computer Vision, 2023, pp. 4195–4205

    work page 2023

  25. [25]

    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. IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY 11

    work page 2022

  26. [26]

    Focal attention for long-range interactions in vision transformers,

    J. Yang, C. Li, P. Zhang, X. Dai, B. Xiao, L. Yuan, and J. Gao, “Focal attention for long-range interactions in vision transformers,” inAdvances in Neural Information Processing Systems, vol. 34, 2021, pp. 30 008– 30 022

    work page 2021

  27. [27]

    Transnext: Robust foveal visual perception for vision transform- ers,

    D. Shi, “Transnext: Robust foveal visual perception for vision transform- ers,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 17 773–17 783

    work page 2024

  28. [28]

    Lsnet: See large, focus small,

    A. Wang, H. Chen, Z. Lin, J. Han, and G. Ding, “Lsnet: See large, focus small,” inProceedings of the Computer Vision and Pattern Recognition Conference, 2025, pp. 9718–9729

    work page 2025

  29. [29]

    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

    work page 2023

  30. [30]

    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,” inProceedings of the 31st ACM International Conference on Multimedia, 2023, pp. 406–416

    work page 2023

  31. [31]

    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, pp. 2249–2262, 2023

    work page 2023

  32. [32]

    Salient object detection in optical remote sensing images driven by transformer,

    G. Li, Z. Bai, Z. Liu, X. Zhang, and H. Ling, “Salient object detection in optical remote sensing images driven by transformer,”IEEE Trans- actions on Image Processing, vol. 32, pp. 5257–5269, 2023

    work page 2023

  33. [33]

    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

  34. [34]

    VSCode: General visual salient and camouflaged object detection with 2d prompt learning,

    Z. Luoet al., “VSCode: General visual salient and camouflaged object detection with 2d prompt learning,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 17 169–17 180

    work page 2024

  35. [35]

    Fantastic animals and where to find them: Segment any marine animal with dual sam,

    P. Zhang, T. Yan, Y . Liu, and H. Lu, “Fantastic animals and where to find them: Segment any marine animal with dual sam,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 2578–2587

    work page 2024

  36. [36]

    Alignment-free rgbt salient object detection: Semantics-guided asymmetric correlation network and a unified benchmark,

    K. Wang, D. Lin, C. Li, Z. Tu, and B. Luo, “Alignment-free rgbt salient object detection: Semantics-guided asymmetric correlation network and a unified benchmark,”IEEE Transactions on Multimedia, vol. 26, pp. 10 692–10 707, 2024

    work page 2024

  37. [37]

    A simple yet effective network based on vision transformer for camouflaged object and salient object detection,

    C. Hao, Z. Yu, X. Liu, J. Xu, H. Yue, and J. Yang, “A simple yet effective network based on vision transformer for camouflaged object and salient object detection,”IEEE Transactions on Image Processing, 2025

    work page 2025

  38. [38]

    Dc-net: Divide-and-conquer for salient object detection,

    J. Zhu, X. Qin, and A. Elsaddik, “Dc-net: Divide-and-conquer for salient object detection,”Pattern Recognition, vol. 157, p. 110903, 2025

    work page 2025

  39. [39]

    Samba: A unified mamba-based framework for general salient object detection,

    J. He, K. Fu, X. Liu, and Q. Zhao, “Samba: A unified mamba-based framework for general salient object detection,” inProceedings of the Computer Vision and Pattern Recognition Conference, 2025, pp. 25 314– 25 324

    work page 2025

  40. [40]

    Perceptual localization and focus refinement network for rgb-d salient object detection,

    J. Han, M. Wang, W. Wu, and X. Jia, “Perceptual localization and focus refinement network for rgb-d salient object detection,”Expert Systems with Applications, vol. 259, p. 125278, 2025

    work page 2025

  41. [41]

    Breaking barriers, localizing saliency: A large-scale benchmark and baseline for condition-constrained salient object detection,

    R. Cong, Z. Chen, H. Fang, S. Kwong, and W. Zhang, “Breaking barriers, localizing saliency: A large-scale benchmark and baseline for condition-constrained salient object detection,”IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 48, no. 4, pp. 4167– 4183, 2026

    work page 2026

  42. [42]

    MAS3K: An open dataset for marine animal segmentation,

    L. Li, E. Rigall, J. Dong, and G. Chen, “MAS3K: An open dataset for marine animal segmentation,” inBenchmarking, Measuring, and Optimizing, ser. Lecture Notes in Computer Science, vol. 12614. Cham: Springer, 2021, pp. 194–212

    work page 2021

  43. [43]

    MASNet: A robust deep marine animal segmentation network,

    Z. Fu, R. Chen, Y . Huang, E. Cheng, X. Ding, and K.-K. Ma, “MASNet: A robust deep marine animal segmentation network,”IEEE Journal of Oceanic Engineering, vol. 49, no. 3, pp. 1104–1115, 2023

    work page 2023

  44. [44]

    Imagenet: A large-scale hierarchical image database,

    J. Deng, W. Dong, R. Socher, L.-J. Li, K. Li, and L. Fei-Fei, “Imagenet: A large-scale hierarchical image database,” in2009 IEEE Conference on Computer Vision and Pattern Recognition. IEEE, 2009, pp. 248–255

    work page 2009

  45. [45]

    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. IEEE, 2012, pp. 733–740

    work page 2012

  46. [46]

    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. IEEE, 2009, pp. 1597–1604

    work page 2009

  47. [47]

    How to evaluate foreground maps?

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

    work page 2014

  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 Twenty-Seventh International Joint Conference on Artificial Intelligence, 2018, pp. 698–704

    work page 2018

  49. [49]

    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,” inProceedings of the IEEE International Conference on Computer Vision, 2017, pp. 4548–4557

    work page 2017

  50. [50]

    Suppress and balance: A simple gated network for salient object detection,

    X. Zhao, Y . Pang, L. Zhang, H. Lu, and L. Zhang, “Suppress and balance: A simple gated network for salient object detection,” in Proceedings of the European Conference on Computer Vision. Springer, 2020, pp. 35–51

    work page 2020

  51. [51]

    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,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2020, pp. 9413–9422

    work page 2020

  52. [52]

    Label decoupling framework for salient object detection,

    J. Wei, S. Wang, Z. Wu, C. Su, Q. Huang, and Q. Tian, “Label decoupling framework for salient object detection,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2020, pp. 13 025–13 034

    work page 2020

  53. [53]

    Visual saliency transformer,

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

    work page 2021

  54. [54]

    EDN: Salient object detection via extremely-downsampled network,

    Y .-H. Wu, Y . Liu, L. Zhang, M.-M. Cheng, and B. Ren, “EDN: Salient object detection via extremely-downsampled network,”IEEE Transactions on Image Processing, vol. 31, pp. 3125–3136, 2022

    work page 2022

  55. [55]

    TRACER: Extreme attention guided salient object tracing network,

    M. S. Lee, W. Shin, and S. W. Han, “TRACER: Extreme attention guided salient object tracing network,” inProceedings of the AAAI Conference on Artificial Intelligence, vol. 36, 2022, pp. 12 993–12 994

    work page 2022

  56. [56]

    Salient object detection via integrity learning,

    M. Zhuge, D.-P. Fan, N. Liu, D. Zhang, D. Xu, and L. Shao, “Salient object detection via integrity learning,”IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 45, no. 3, pp. 3738–3752, 2022

    work page 2022

  57. [57]

    Towards a complete and detail-preserved salient object detection,

    Y . K. Yun and W. Lin, “Towards a complete and detail-preserved salient object detection,”IEEE Transactions on Multimedia, vol. 26, pp. 4667– 4680, 2024

    work page 2024

  58. [58]

    Boosting broader receptive fields for salient object detection,

    M. Ma, C. Xia, C. Xie, X. Chen, and J. Li, “Boosting broader receptive fields for salient object detection,”IEEE Transactions on Image Processing, vol. 32, pp. 1026–1038, 2023

    work page 2023

  59. [59]

    Size-invariance matters: Rethinking metrics and losses for imbalanced multi-object salient object detection,

    F. Liet al., “Size-invariance matters: Rethinking metrics and losses for imbalanced multi-object salient object detection,” inProceedings of the International Conference on Machine Learning, 2024, pp. 28 989– 29 021

    work page 2024

  60. [60]

    Multi-scale and detail-enhanced segment anything model for salient object detection,

    S. Gao, P. Zhang, T. Yan, and H. Lu, “Multi-scale and detail-enhanced segment anything model for salient object detection,” inProceedings of the 32nd ACM International Conference on Multimedia, 2024, pp. 9894–9903

    work page 2024

  61. [61]

    Modeling aleatoric uncertainty for camouflaged object detection,

    J. Liu, J. Zhang, and N. Barnes, “Modeling aleatoric uncertainty for camouflaged object detection,” inProceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, 2022, pp. 1445–1454

    work page 2022

  62. [62]

    Zoom in and out: A mixed-scale triplet network for camouflaged object detection,

    Y . Pang, X. Zhao, T.-Z. Xiang, L. Zhang, and H. Lu, “Zoom in and out: A mixed-scale triplet network for camouflaged object detection,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2022, pp. 2160–2170

    work page 2022

  63. [63]

    Rethinking semantic segmen- tation from a sequence-to-sequence perspective with transformers,

    S. Zheng, J. Lu, H. Zhao, X. Zhu, Z. Luo, Y . Wang, Y . Fu, J. Feng, T. Xiang, P. H. S. Torr, and L. Zhang, “Rethinking semantic segmen- tation from a sequence-to-sequence perspective with transformers,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2021, pp. 6881–6890

    work page 2021

  64. [64]

    H2Former: An effi- cient hierarchical hybrid transformer for medical image segmentation,

    A. He, K. Wang, T. Li, C. Du, S. Xia, and H. Fu, “H2Former: An effi- cient hierarchical hybrid transformer for medical image segmentation,” IEEE Transactions on Medical Imaging, vol. 42, no. 9, pp. 2763–2775, 2023

    work page 2023

  65. [65]

    Segment anything,

    A. Kirillov, E. Mintun, N. Ravi, H. Mao, C. Rolland, L. Gustafson, T. Xiao, S. Whitehead, A. C. Berg, W.-Y . Lo, P. Doll´ar, and R. Girshick, “Segment anything,” inProceedings of the IEEE/CVF International Conference on Computer Vision, 2023, pp. 4015–4026

    work page 2023

  66. [66]

    SAM-Adapter: Adapting segment anything in underperformed scenes,

    T. Chen, L. Zhu, C. Deng, R. Cao, Y . Wang, S. Zhang, Z. Li, L. Sun, Y . Zang, and P. Mao, “SAM-Adapter: Adapting segment anything in underperformed scenes,” inProceedings of the IEEE/CVF International Conference on Computer Vision Workshops, 2023, pp. 3367–3375

    work page 2023

  67. [67]

    Cross-modal fusion and progressive decoding network for rgb-d salient object detection,

    X. Hu, F. Sun, J. Sun, F. Wang, and H. Li, “Cross-modal fusion and progressive decoding network for rgb-d salient object detection,” International Journal of Computer Vision, vol. 132, no. 8, pp. 3067– 3085, 2024

    work page 2024

  68. [68]

    Mas-sam: segment any marine animal with aggregated features,

    T. Yan, Z. Wan, X. Deng, P. Zhang, Y . Liu, and H. Lu, “Mas-sam: segment any marine animal with aggregated features,” inProceedings of the Thirty-Third International Joint Conference on Artificial Intelligence, 2024, pp. 6886–6894

    work page 2024