Pith. sign in

REVIEW 4 major objections 7 minor 103 references

Excavate the potential of Single-Scale Features: A Decomposition Network for Water-Related Optical Image Enhancement

T0 review · 4 major / 7 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read A decomposition network that works entirely at single scale can match or beat multi-scale architectures for underwater image enhancement, reporting top PSNR/SSIM scores on UIEB and EUVP with only 1.03M parameters.

desk verdict An efficient new architecture whose central 'single-scale' claim is undercut by downsampling inside its own PFDB; worth refereeing, but the conclusion needs to be reframed and verified. read the letter →

arxiv 2508.04123 v1 pith:MH726YJZ submitted 2025-08-06 cs.CV eess.IV

Zheng Cheng , Wenri Wang , Guangyong Chen , Yakun Ju , Yihua Cheng , Zhisong Liu , Yanda Meng , Jintao Song This is my paper
classification cs.CVeess.IV
keywords underwaterimageenhancementsingle-scalenetworkfeaturedecompositionadaptivesparsetransformermulti-scaleextractionrestorationefficientmodel
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

Underwater images are degraded by light absorption and scattering, and nearly every modern enhancement network assumes that fusing features across multiple resolutions is essential for good restoration. This paper argues the opposite: a well-designed network operating at a single resolution can match or beat multi-scale architectures. To show it, the authors build SSD-Net, a 1.03M-parameter network that splits features into a clean branch and a degradation branch, refines them through parallel decomposition and bidirectional communication blocks, and reports leading PSNR/SSIM scores on the UIEB and EUVP benchmarks. In a structural ablation, converting SSD-Net into U-Net-like, progressive-resolution, and parallel-branch forms adds 50–160% more parameters while gaining little or no quality. If the claim holds, the field's default reliance on multi-scale fusion is a costly habit, not a necessity.

What carries the argument

The load-bearing mechanism is the asymmetric dual-branch feature decomposition formed by two stacked modules. The Parallel Feature Decomposition Block (PFDB) splits the single-resolution representation into a degradation stream modeled by an Adaptive Sparse Transformer and a clean-content stream modeled by lightweight CNN channel attention. The Bidirectional Feature Communication Block (BFCB) then exchanges information between streams: sigmoid gates from $1\times1$ convolutions plus learnable scaling produce coupled residuals, and each branch is updated by subtracting its own redundancy and adding the other's useful signal. The third piece, Adaptive Sparse Attention, blends a dense softmax b

What would settle it

Retrain the Table III structural variants (U-Net, PMR, PMB) and SSD-Net on EUVP under identical hyperparameters, loss, and compute budget; if any multi-scale variant reaches or exceeds roughly 29.84 dB once equally tuned, the single-scale advantage claim fails. A second check: remove the down/up-sampling that Eq. (8) places around the Transformer blocks — if PSNR drops materially, the reported gains come from multi-resolution processing rather than single-scale features.

Watch

Extended reading notes

Core claim

Central claim: high-quality underwater image reconstruction does not need multi-scale feature fusion. SSD-Net works at one resolution, splitting the input into clean features $F_c$ and degradation features $F_d$ via the Parallel Feature Decomposition Block (PFDB) — a CNN channel-attention branch plus a Transformer branch fusing dense (softmax) and sparse (ReLU) attention — and the Bidirectional Feature Communication Block (BFCB) for cross-stream residual exchange. It reports $24.90$ dB PSNR / $0.924$ SSIM on UIEB and $29.84$ dB / $0.917$ on EUVP, ahead of the compared methods, with $1.03$M parameters. Turned into U-Net, progressive multi-resolution, and parallel multi-branch variants, the mu

Load-bearing premise

The claim stands on the premise that SSD-Net is genuinely single-scale and that the structural ablation is fair — but the paper's own Eq. (8) puts down-sampling and up-sampling around the Transformer blocks, so if those operations count as multi-scale processing, the experiments no longer compare single-scale against multi-scale features; the fairness of the Table III variants also depends on transformation details the paper does not give.

Editorial extensions

If this is right

  • On UIEB and EUVP, SSD-Net reports the best scores among the compared methods, so a single-scale decomposition design can serve as a drop-in alternative to multi-scale backbones for underwater enhancement.
  • Table III shows the multi-scale restructurings add 49.9%–160.8% parameters without a consistent quality gain, implying much of the cost of current UIE networks goes into redundant cross-scale features.
  • The unpaired-set results on UIEB60, U45, and UCCS indicate the single-scale design generalizes across color casts, turbidity, and low-light conditions.
  • With 1.03M parameters, the efficiency benefit is direct: comparable or better restoration at a fraction of the model size of competing learned methods, which matters for deployment on underwater vehicles and embedded cameras.

Reading between the lines

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

  • The paper says SSD-Net uses no downsampling, but its own Eq. (8) sandwiches the Transformer blocks in down-sampling and up-sampling operations; if those count as multi-scale processing, the defensible reading is not that scale is useless but that the decomposition-and-cross-talk design, rather than scale, is what drives the gains.
  • The clean/degradation split is not underwater-specific: the same PFDB–BFCB design could be tested on dehazing, deraining, and low-light enhancement, where multi-scale backbones dominate; a positive result there would widen the claim beyond water-related optics.
  • The paper reports only parameter counts, not FLOPs or latency; an inference-time comparison against the 1.78–1.84M-parameter baselines on embedded hardware would settle whether the practical efficiency advantage is as large as the parameter savings suggest.
  • Because the sparse (ReLU) attention branch is fused adaptively with the dense branch, one testable consequence is that attention sparsity — not global modeling per se — carries the benefit; ablating each branch separately would isolate which mechanism the gain comes from.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 7 minor

Summary. The paper proposes SSD-Net for underwater image enhancement, claiming that single-scale feature extraction can match or surpass multi-scale methods. The network uses a dual-branch decomposition into clean and degradation features, with two core modules: PFDB (CNN branch plus Transformer branch with adaptive sparse attention) and BFCB (bidirectional residual exchange). Losses include SSIM and L1 on both the clean output and the reconstructed input. Experiments on UIEB, EUVP, and no-reference datasets report state-of-the-art PSNR/SSIM with 1.03M parameters. Ablations include component ablation, architectural variants (U-Net, PMR, PMB), and block count.

Significance. If the single-scale claim were established, the paper would provide a useful counterpoint to the prevalence of multi-scale designs in image restoration, potentially enabling more efficient models. The architecture also introduces a plausible decomposition mechanism and an adaptive sparse attention variant. Strengths: the manuscript reports thorough quantitative comparisons on multiple datasets, a component ablation, a structural ablation, and a parameter/performance trade-off analysis; the proposed modules are concretely specified and the loss is clearly defined. No code or trained models are released, however, and the central conceptual claim is undermined by an internal contradiction about down/up-sampling within PFDB.

major comments (4)
  1. [Sec. III-A and Sec. III-C, Eq. (8)] The central claim that SSD-Net is 'single-scale' is contradicted by the PFDB description. Sec. III-A states the network 'does not adopt a U-Net-like encoder-decoder structure nor employs pooling or downsampling operations,' yet Sec. III-C states the Transformer branch 'sandwich(es) the Transformer blocks with down-sampling and up-sampling operations,' and Eq. (8) explicitly contains down- and up-sampling arrows. Downsampling followed by upsampling creates multiple spatial resolutions inside the network, so by the paper's own definition the network is not single-scale. This is load-bearing because the abstract and the structural ablation (Table III) rest on the premise that SSD-Net is genuinely single-scale. A version of SSD-Net with no down/up-sampling must be ablated, or the claim must be reworded to exclude the transformer branch, otherwise the comparison does not establish that multi-
  2. [Table III and Sec. IV-C.2] The structural ablation is not sufficient to support the 'single-scale matches multi-scale' conclusion. The three 'multi-scale' variants (U-Net, PMR, PMB) are described only by name; the paper does not specify which modules are duplicated, how down/up-sampling was inserted, what channels/resolutions were used, whether the PFDB's internal down/up-sampling was removed in the SSN baseline, or whether training budgets were matched. Since the parameter counts differ substantially (1.026M to 2.608M), the performance comparison conflates architecture scale with capacity and implementation effort. Without detailed configurations and controlled capacity/computation, the conclusion that the performance difference is due specifically to scale is not established.
  3. [Table I and Sec. IV.B] The reported quantitative gains lack any measure of variability or reproducibility. No code, no pre-trained models, no random seed information, and no error bars are provided. The performance differences over the second-best methods are small on some metrics (e.g., 0.924 vs 0.909 SSIM on UIEB; 29.84 vs 28.00 PSNR on EUVP), and a single-run comparison cannot rule out training stochasticity. At minimum, the authors should report multiple seeds with means and standard deviations, and ideally release code/model checkpoints to allow independent verification.
  4. [Sec. III-E, Eq. (15)] The loss function is under-specified. Eq. (13) defines an SSIM loss with exponents α and β, while Eq. (15) re-uses α and β as weights for the reconstruction terms L_ssim(X',X) and L1(X',X). The same symbols appear to denote different quantities, and the manuscript never states the values of the exponents in Eq. (13) or whether the 'L1' in Eq. (14) is used with the SSIM term as a multiplicative product (as in the phrase 'SSIM-based loss function is defined as: L_ssim = (1 - ...)^α (...)^β') or as a separate additive term. This ambiguity makes the exact training objective impossible to reproduce.
minor comments (7)
  1. [Sec. III-C, text after Eq. (9)] The subscripts in the notation appear inconsistent: the text defines F_d as clear features and F_c as degradation features in Sec. III-B, but the PFDB description refers to 'Transformer-driven Degradation Feature Branch' taking F_d and 'CNN-driven Clear Feature Branch' taking F_c, and Fig. 3 labels appear to swap the roles. Please standardize the notation and figure labels.
  2. [Sec. III-D, Eqs. (10)-(12)] There are typographical errors in the equations: G^n-1_d->c and the residual formulas contain mismatched brackets and subscripts (e.g., [F^{n-1}_d) and res_d->c = [F^{n-1}_d * ...; the second uses G^{n-1}_{d->d} rather than G^{n-1}_{d->c}). Please correct the formulas.
  3. [Table II caption and rows] The component ablation is hard to interpret because the table only shows checkmarks without a legend for how the 'conventional Transformer' or 'dual-branch convolutional network' baselines were constructed. Add the missing legend and describe the baseline configurations in the text.
  4. [Sec. IV.C.2, Table III caption] The caption contains a garbled placeholder: 'Table III able xxx compares... Here, "3S" denotes Single-Scale Structure.' Please fix the caption and the typo 'Efficient' in the author bios if applicable.
  5. [Sec. IV.A and Table IV] The ablation on N and M reports PSNR values that differ by 0.01-0.02 dB for several configurations (e.g., N=4,M=6 vs N=6,M=6), likely within noise. Please add error bars or state explicitly that these differences are not significant.
  6. [Fig. 6 caption] The qualitative comparison on EUVP is labeled 'LSUI dataset' in the caption, which is inconsistent with the text (EUVP). Please correct.
  7. [Overall] Several minor formatting issues: 'PFCB' should be 'BFCB' in Sec. IV.A.1; the references list includes a paper with a garbled author string 'Z. L. W. H. Y. Y. W. L. X.-B. Li, H.-F' [28]; 'IJCNN' should be 'IJCNN' [66]. Please proofread.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the single-scale labeling contradiction is a correctness/validity concern, not a derivation that reduces to its inputs.

full rationale

The paper's argument is empirical: it reports PSNR/SSIM and parameter counts on external benchmarks and ablations. No fitted parameter is renamed as a prediction, no uniqueness theorem is imported from the authors' prior work, and no load-bearing step is justified only by a self-citation. The decomposition X' = X_c + X_d (Sec. III-B, Eq. 15) does define X_d as the reconstruction residual, so calling it a 'degradation residual' is a label rather than a derived physical quantity, but this does not by itself make any reported performance result equivalent to the model's inputs. The reader's noted contradiction is real and important: Sec. III-A says SSD-Net 'does not adopt a U-Net-like encoder-decoder structure nor employs pooling or downsampling operations for feature compression,' while Sec. III-C says the PFDB 'sandwich(es) the Transformer blocks with down-sampling and up-sampling operations' and Eq. 8 contains down/up arrows. This undermines the construct validity of the paper's claim that SSD-Net is genuinely single-scale and weakens the interpretation of the Table III ablation. However, an internal labeling inconsistency is a correctness/validity problem, not a circular derivation: the reported comparisons against external SOTA methods remain independent evidence, and the paper does not derive its central claim from the label itself. Under the strict definition of circularity used here, no step reduces by construction to its own input.

Assumptions & free parameters 3 free parameters · 3 assumptions · 1 invented entities

The paper rests on standard supervised learning assumptions, a small number of hand-chosen hyperparameters, and a conceptual decomposition that is not independently validated. The fair-comparison assumption for the structural ablations is the most fragile item.

free parameters (3)
  • Loss weights alpha and beta in Eq. (15) = 0.2 each
    Empirically set hyperparameters weighting SSIM and L1 losses on the reconstructed image; not derived or swept.
  • Number of cascaded PFDB/BFCB blocks N = 4
    Chosen from ablation on EUVP (Table IV) as the saturation point.
  • Number of AST modules M per PFDB = 4
    Chosen from ablation on EUVP (Table IV) as the saturation point.
assumptions (3)
  • domain assumption The UIEB and EUVP reference images Y are treated as ground-truth clean images
    The whole supervised training and evaluation assumes paired references accurately represent desired enhancement; stated in Sec. IV-A.3.
  • domain assumption The transformation of SSD-Net into U-Net, PMR, and PMB variants retains the core module design fairly
    The key structural comparison in Table III relies on this fairness, but implementation details are not provided; Sec. IV-C.2.
  • ad hoc to paper Replacing Softmax with a parallel ReLU branch improves attention sparsity without hurting trainability
    This design choice is motivated only by reasoning about Softmax amplification, not by a formal guarantee or external benchmark; Sec. III-C, Eq. (6).
invented entities (1)
  • Clean and degradation layer decomposition
    purpose: To motivate the dual-branch architecture and residual prediction where one branch encodes scene content and the other encodes medium-induced interference.
    The paper does not provide evidence that these internal branches correspond to physically meaningful layers; they are learned feature representations.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Excavate the potential of Single-Scale Features: A Decomposition Network for Water-Related Optical Image Enhancement." pith.science (2026). https://pith.science/paper/MH726YJZ

@misc{pith2026250804123,
  author       = {Pith},
  title        = {Pith review of: Excavate the potential of Single-Scale Features: A Decomposition Network for Water-Related Optical Image Enhancement},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/MH726YJZ}},
  note         = {Machine review of arXiv:2508.04123}
}
read the original abstract

Underwater image enhancement (UIE) techniques aim to improve visual quality of images captured in aquatic environments by addressing degradation issues caused by light absorption and scattering effects, including color distortion, blurring, and low contrast. Current mainstream solutions predominantly employ multi-scale feature extraction (MSFE) mechanisms to enhance reconstruction quality through multi-resolution feature fusion. However, our extensive experiments demonstrate that high-quality image reconstruction does not necessarily rely on multi-scale feature fusion. Contrary to popular belief, our experiments show that single-scale feature extraction alone can match or surpass the performance of multi-scale methods, significantly reducing complexity. To comprehensively explore single-scale feature potential in underwater enhancement, we propose an innovative Single-Scale Decomposition Network (SSD-Net). This architecture introduces an asymmetrical decomposition mechanism that disentangles input image into clean layer along with degradation layer. The former contains scene-intrinsic information and the latter encodes medium-induced interference. It uniquely combines CNN's local feature extraction capabilities with Transformer's global modeling strengths through two core modules: 1) Parallel Feature Decomposition Block (PFDB), implementing dual-branch feature space decoupling via efficient attention operations and adaptive sparse transformer; 2) Bidirectional Feature Communication Block (BFCB), enabling cross-layer residual interactions for complementary feature mining and fusion. This synergistic design preserves feature decomposition independence while establishing dynamic cross-layer information pathways, effectively enhancing degradation decoupling capacity.

Figures

Figures reproduced from arXiv: 2508.04123 by the authors.

Figure 1
Figure 1. Comparison of the number of model parameters and corresponding [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Comparison of several common image restoration network archi [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. An overview of the proposed SSD-Net. The network is designed as a single-scale architecture, consisting of cascaded PFDB and BFCB modules. It [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: illustrates the proposed Adaptive Sparse Attention mechanism. The [PITH_FULL_IMAGE:figures/full_fig_p006_4.png]
Figure 5
Figure 5. Figure 5: Qualitative comparison on the UIEB dataset. Please zoom in for better visualization of fine details. The bottom-right corner of each image shows the [PITH_FULL_IMAGE:figures/full_fig_p010_5.png]
Figure 6
Figure 6. Figure 6: Qualitative comparison on the LSUI dataset. Please zoom in to inspect fine details. The PSNR and SSIM values are shown at the bottom-right corner [PITH_FULL_IMAGE:figures/full_fig_p010_6.png]
Figure 7
Figure 7. Figure 7: Qualitative comparison on the U45, UCCS, and UIEB datasets. Please zoom in to examine fine details. The UCIQE score for each image is shown in [PITH_FULL_IMAGE:figures/full_fig_p011_7.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

103 extracted references · 69 canonical work pages

  1. [1]

    Computer Modeling and the Design of Optimal Underwater Imaging Systems,

    J. S. Jaffe, “Computer Modeling and the Design of Optimal Underwater Imaging Systems,” IEEE Journal of Oceanic Engineering, vol. 15, no. 2, pp. 101–111, 1990

    1990

  2. [2]

    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

    2019

  3. [3]

    Underwater Image Enhancement by Wavelength Compensation and Dehazing,

    J. Y. Chiang and Y.-C. Chen, “Underwater Image Enhancement by Wavelength Compensation and Dehazing,”IEEE Transactions on Im- age Processing, vol. 21, no. 4, pp. 1756–1769, 2011

    2011

  4. [4]

    Underwater Single Image Color Restoration using Haze-lines and a New Quantitative Dataset,

    D. Berman, D. Levy, S. Avidan, and T. Treibitz, “Underwater Single Image Color Restoration using Haze-lines and a New Quantitative Dataset,” IEEE Transactions on Pattern Analysis and Machine Intelli- gence, vol. 43, no. 8, pp. 2822–2837, 2020

    2020

  5. [5]

    A Retinex-based Enhancing Approach for Single Underwater Image,

    X. Fu, P. Zhuang, Y. Huang, Y. Liao, X.-P. Zhang, and X. Ding, “A Retinex-based Enhancing Approach for Single Underwater Image,” in Proceedings of the International Conference on Image Processing. IEEE, 2014, pp. 4572–4576

    2014

  6. [6]

    UIECˆ 2-Net: CNN-based Underwater Image Enhancement using Wwo Color Space,

    Y. Wang, J. Guo, H. Gao, and H. Yue, “UIECˆ 2-Net: CNN-based Underwater Image Enhancement using Wwo Color Space,” Signal Processing: Image Communication, vol. 96, pp. 116250–116261, 2021

    2021

  7. [7]

    IACC: Cross-Illumination Awareness and Color Correction for Underwater Images Under Mixed Natural and Artificial Lighting,

    J. Zhou, Q. Gai, D. Zhang, K.-M. Lam, W. Zhang, and X. Fu, “IACC: Cross-Illumination Awareness and Color Correction for Underwater Images Under Mixed Natural and Artificial Lighting,” IEEE Trans- actions on Geoscience and Remote Sensing, vol. 62, pp. 1–15, 2024

    2024

  8. [8]

    Robust Underwater Image Enhancement with Cascaded Multi-level Sub-networks and Triple Attention Mechanism,

    D. Zhang, C. Wu, J. Zhou, W. Zhang, Z. Lin, K. Polat, and F. Alenezi, “Robust Underwater Image Enhancement with Cascaded Multi-level Sub-networks and Triple Attention Mechanism,”Neural Networks, vol. 169, pp. 685–697, 2024

    2024

Show all 103 references
  1. [9]

    Wavelet-based fourier infor- mation interaction with frequency diffusion adjustment for underwater image restoration,

    C. Zhao, W. Cai, C. Dong, and C. Hu, “Wavelet-based fourier infor- mation interaction with frequency diffusion adjustment for underwater image restoration,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 8281–8291

    2024

  2. [10]

    Synergistic multi- scale detail refinement via intrinsic supervision for underwater image enhancement,

    D. Zhang, J. Zhou, C. Guo, W. Zhang, and C. Li, “Synergistic multi- scale detail refinement via intrinsic supervision for underwater image enhancement,” in Proceedings of the AAAI conference on artificial intelligence, vol. 38, no. 7, 2024, pp. 7033–7041

    2024

  3. [11]

    Task-friendly underwater image enhancement for machine vision applications,

    M. Yu, L. Shen, Z. Wang, and X. Hua, “Task-friendly underwater image enhancement for machine vision applications,”IEEE Transactions on Geoscience and Remote Sensing, vol. 62, pp. 1–14, 2023

    2023

  4. [12]

    Fast underwater image enhance- ment for improved visual perception,

    M. J. Islam, Y. Xia, and J. Sattar, “Fast underwater image enhance- ment for improved visual perception,”IEEE Robotics and Automation Letters, vol. 5, no. 2, pp. 3227–3234, 2020

    2020

  5. [13]

    U-shape transformer for underwater image enhancement,

    L. Peng, C. Zhu, and L. Bian, “U-shape transformer for underwater image enhancement,”IEEE Transactions on Image Processing, vol. 32, pp. 3066–3079, 2023

    2023

  6. [14]

    Contrastive Semi- supervised Learning for Underwater Image Restoration via Reliable Bank,

    S. Huang, K. Wang, H. Liu, J. Chen, and Y. Li, “Contrastive Semi- supervised Learning for Underwater Image Restoration via Reliable Bank,” in Proceedings of the Conference on Computer Vision and Pattern Recognition. IEEE, 2023, pp. 18145–18155

    2023

  7. [15]

    Twin Adversarial Contrastive Learning for Underwater Image Enhancement and Beyond,

    R. Liu, Z. Jiang, S. Yang, and X. Fan, “Twin Adversarial Contrastive Learning for Underwater Image Enhancement and Beyond,” IEEE Transactions on Image Processing, vol. 31, pp. 4922–4936, 2022

    2022

  8. [16]

    Underwater image enhancement via medium transmission-guided multi-color space embedding,

    C. Li, S. Anwar, J. Hou, R. Cong, C. Guo, and W. Ren, “Underwater image enhancement via medium transmission-guided multi-color space embedding,” IEEE Transactions on Image Processing, vol. 30, pp. 4985–5000, 2021

    2021

  9. [17]

    Underwater image enhancement based on deep learning and image formation model,

    X. Chen, P. Zhang, L. Quan, C. Yi, and C. Lu, “Underwater image enhancement based on deep learning and image formation model,” arXiv preprint arXiv:2101.00991, 2021

    2021 arXiv

  10. [18]

    Towards real-time advancement of underwater visual quality with gan,

    X. Chen, J. Yu, S. Kong, Z. Wu, X. Fang, and L. Wen, “Towards real-time advancement of underwater visual quality with gan,”IEEE Transactions on Industrial Electronics, vol. 66, no. 12, pp. 9350–9359, 2019. JOURNAL OF LATEX CLASS FILES, VOL. 14, NO. 8, AUGUST 2021 12

    2019

  11. [19]

    GUDCP: Gen- eralization of underwater dark channel prior for underwater image restoration,

    Z. Liang, X. Ding, Y. Wang, X. Yan, and X. Fu, “GUDCP: Gen- eralization of underwater dark channel prior for underwater image restoration,” IEEE Transactions on Circuits and Systems for Video Technology, vol. 32, no. 7, pp. 4879–4884, 2021

    2021

  12. [20]

    Transmission estimation in underwater single images,

    P. Drews, E. Nascimento, F. Moraes, S. Botelho, and M. Campos, “Transmission estimation in underwater single images,” inProceedings of the IEEE international conference on computer vision workshops, 2013, pp. 825–830

    2013

  13. [21]

    Underwater image restoration based on image blurriness and light absorption,

    Y.-T. Peng and P. C. Cosman, “Underwater image restoration based on image blurriness and light absorption,” IEEE transactions on image processing, vol. 26, no. 4, pp. 1579–1594, 2017

    2017

  14. [22]

    Single underwater image restoration using adaptive attenuation-curve prior,

    Y. Wang, H. Liu, and L.-P. Chau, “Single underwater image restoration using adaptive attenuation-curve prior,”IEEE Transactions on Circuits and Systems I: Regular Papers, vol. 65, no. 3, pp. 992–1002, 2017

    2017

  15. [23]

    Single underwater image restoration by decomposing curves of attenuating color,

    C. Dai, M. Lin, X. Wu, Z. Wang, and Z. Guan, “Single underwater image restoration by decomposing curves of attenuating color,”Optics & Laser Technology, vol. 123, p. 105947, 2020

    2020

  16. [24]

    Underwater Image Enhancement Method via Multi-Interval Subhistogram Perspective Equalization,

    J. Zhou, L. Pang, D. Zhang, and W. Zhang, “Underwater Image Enhancement Method via Multi-Interval Subhistogram Perspective Equalization,” IEEE Journal of Oceanic Engineering, vol. 48, no. 2, pp. 474–488, 2023

    2023

  17. [25]

    Adaptive Histogram Equalization and its Variations,

    S. M. Pizer, E. P. Amburn, J. D. Austin, R. Cromartie, A. Geselowitz, T. Greer, B. ter Haar Romeny, J. B. Zimmerman, and K. Zuiderveld, “Adaptive Histogram Equalization and its Variations,”Computer Vi- sion, Graphics, and Image Processing, vol. 39, no. 3, pp. 355–368, 1987

    1987

  18. [26]

    Mixture Contrast Limited Adaptive Histogram Equalization for Underwater Image Enhancement,

    M. S. Hitam, E. A. Awalludin, W. N. J. H. W. Yussof, and Z. Ba- chok, “Mixture Contrast Limited Adaptive Histogram Equalization for Underwater Image Enhancement,” inProceedings of the International Conference on Computer Applications Technology. IEEE, 2013, pp. 1–5

    2013

  19. [27]

    Under- water Image Enhancement by Dehazing with Minimum Information Loss and Histogram Distribution Prior,

    C.-Y. Li, J.-C. Guo, R.-M. Cong, Y.-W. Pang, and B. Wang, “Under- water Image Enhancement by Dehazing with Minimum Information Loss and Histogram Distribution Prior,”IEEE Transactions on Image Processing, vol. 25, no. 12, pp. 5664–5677, 2016

    2016

  20. [28]

    Polarimetric Image Recovery Method Combining Histogram Stretching for Underwater Imaging,

    Z. L. W. H. Y. Y. W. L. X.-B. Li, H.-F and T. Liu, “Polarimetric Image Recovery Method Combining Histogram Stretching for Underwater Imaging,” Scientific reports, vol. 8, no. 1, p. 12430, 2018

    2018

  21. [29]

    Color balance and fusion for underwater image enhancement,

    C. O. Ancuti, C. Ancuti, C. De Vleeschouwer, and P. Bekaert, “Color balance and fusion for underwater image enhancement,”IEEE Trans- actions on image processing, vol. 27, no. 1, pp. 379–393, 2017

    2017

  22. [30]

    Underwater image enhancement based on color balance and multi-scale fusion,

    Q. Chen, Z. Zhang, and G. Li, “Underwater image enhancement based on color balance and multi-scale fusion,” IEEE Photonics Journal, vol. 14, no. 6, pp. 1–10, 2022

    2022

  23. [31]

    A new color correction method for underwater imaging,

    G. Bianco, M. Muzzupappa, F. Bruno, R. Garcia, and L. Neumann, “A new color correction method for underwater imaging,”The Inter- national Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, vol. 40, pp. 25–32, 2015

    2015

  24. [32]

    Joint Iterative Color Correction and Dehazing for Underwater Image Enhancement,

    K. Wang, L. Shen, Y. Lin, M. Li, and Q. Zhao, “Joint Iterative Color Correction and Dehazing for Underwater Image Enhancement,”IEEE Robotics and Automation Letters, vol. 6, no. 3, pp. 5121–5128, 2021

    2021

  25. [33]

    Multi-view underwater image enhancement method via embedded fusion mechanism,

    J. Zhou, J. Sun, W. Zhang, and Z. Lin, “Multi-view underwater image enhancement method via embedded fusion mechanism,”Engineering applications of artificial intelligence, vol. 121, p. 105946, 2023

    2023

  26. [34]

    Underwater image enhancement method via multi-feature prior fusion,

    J. Zhou, D. Zhang, and W. Zhang, “Underwater image enhancement method via multi-feature prior fusion,”Applied Intelligence, vol. 52, no. 14, pp. 16435–16457, 2022

    2022

  27. [35]

    Underwater image enhancement via weighted wavelet visual percep- tion fusion,

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

    2023

  28. [36]

    Underwater Scene Prior Inspired Deep Underwater Image and Video Enhancement,

    C. Li, S. Anwar, and F. Porikli, “Underwater Scene Prior Inspired Deep Underwater Image and Video Enhancement,” Pattern Recognition, vol. 98, pp. 107038–107049, 2020

    2020

  29. [37]

    Underwater Image Enhancement via Minimal Color Loss and Lo- cally Adaptive Contrast Enhancement,

    W. Zhang, P. Zhuang, H.-H. Sun, G. Li, S. Kwong, and C. Li, “Underwater Image Enhancement via Minimal Color Loss and Lo- cally Adaptive Contrast Enhancement,”IEEE Transactions on Image Processing, vol. 31, pp. 3997–4010, 2022

    2022

  30. [38]

    Ca- gan: Class-condition attention gan for underwater image enhancement,

    J. Wang, P. Li, J. Deng, Y. Du, J. Zhuang, P. Liang, and P. Liu, “Ca- gan: Class-condition attention gan for underwater image enhancement,” IEEE access, vol. 8, pp. 130719–130728, 2020

    2020

  31. [39]

    Pugan: Physical model-guided underwater image enhance- ment using gan with dual-discriminators,

    R. Cong, W. Yang, W. Zhang, C. Li, C.-L. Guo, Q. Huang, and S. Kwong, “Pugan: Physical model-guided underwater image enhance- ment using gan with dual-discriminators,”IEEE Transactions on Image Processing, vol. 32, pp. 4472–4485, 2023

    2023

  32. [40]

    Uw-gan: Single-image depth estimation and image enhancement for underwater images,

    P. Hambarde, S. Murala, and A. Dhall, “Uw-gan: Single-image depth estimation and image enhancement for underwater images,” IEEE Transactions on Instrumentation and Measurement, vol. 70, pp. 1–12, 2021

    2021

  33. [41]

    Enhancing underwater imagery using generative adversarial networks,

    C. Fabbri, M. J. Islam, and J. Sattar, “Enhancing underwater imagery using generative adversarial networks,” in 2018 IEEE international conference on robotics and automation (ICRA). IEEE, 2018, pp. 7159–7165

    2018

  34. [42]

    An Underwater Image Enhancement Benchmark Dataset and Beyond,

    C. Li, C. Guo, W. Ren, R. Cong, J. Hou, S. Kwong, and D. Tao, “An Underwater Image Enhancement Benchmark Dataset and Beyond,” IEEE Transactions on Image Processing, vol. 29, pp. 4376–4389, 2019

    2019

  35. [43]

    Underwater image restoration via polymorphic large kernel cnns,

    X. Guo, Y. Dong, X. Chen, W. Chen, Z. Li, F. Zheng, and C.-M. Pun, “Underwater image restoration via polymorphic large kernel cnns,” in ICASSP 2025-2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, 2025, pp. 1–5

    2025

  36. [44]

    Cdf-uie: Leveraging cross-domain fusion for underwater image enhancement,

    H. Zhang, H. Xu, X. Yu, X. Zhang, X. Gao, and C. Wu, “Cdf-uie: Leveraging cross-domain fusion for underwater image enhancement,” IEEE Transactions on Geoscience and Remote Sensing, 2025

    2025

  37. [45]

    Phaseformer: Phase-based attention mechanism for underwater image restoration and beyond,

    M. R. Khan, A. Negi, A. Kulkarni, S. S. Phutke, S. K. Vipparthi, and S. Murala, “Phaseformer: Phase-based attention mechanism for underwater image restoration and beyond,” IEEE, pp. 9618–9629, 2025

    2025

  38. [46]

    Towards progressive multi-frequency representation for image warping,

    J. Xiao, Z. Lyu, C. Zhang, Y. Ju, C. Shui, and K.-M. Lam, “Towards progressive multi-frequency representation for image warping,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 2995–3004

    2024

  39. [47]

    A deep- shallow and global–local multi-feature fusion network for photometric stereo,

    Y. Liu, Y. Ju, M. Jian, F. Gao, Y. Rao, Y. Hu, and J. Dong, “A deep- shallow and global–local multi-feature fusion network for photometric stereo,” Image and Vision Computing, vol. 118, p. 104368, 2022

    2022

  40. [48]

    Pay attention to devils: A photometric stereo network for better details

    Y. Ju, K.-M. Lam, Y. Chen, L. Qi, J. Donget al., “Pay attention to devils: A photometric stereo network for better details.” International Joint Conference on Artificial Intelligence, 2020

    2020

  41. [49]

    Esti- mating high-resolution surface normals via low-resolution photometric stereo images,

    Y. Ju, M. Jian, C. Wang, C. Zhang, J. Dong, and K.-M. Lam, “Esti- mating high-resolution surface normals via low-resolution photometric stereo images,”IEEE Transactions on Circuits and Systems for Video Technology, vol. 34, no. 4, pp. 2512–2524, 2023

    2023

  42. [50]

    Gr-psn: Learning to estimate surface normal and reconstruct photometric stereo images,

    Y. Ju, B. Shi, Y. Chen, H. Zhou, J. Dong, and K.-M. Lam, “Gr-psn: Learning to estimate surface normal and reconstruct photometric stereo images,”IEEE Transactions on Visualization and Computer Graphics, vol. 30, no. 9, pp. 6192–6207, 2023

    2023

  43. [51]

    Efficient inductive vision transformer for oriented object detection in remote sensing imagery,

    C. Zhang, J. Su, Y. Ju, K.-M. Lam, and Q. Wang, “Efficient inductive vision transformer for oriented object detection in remote sensing imagery,” IEEE Transactions on Geoscience and Remote Sensing , vol. 61, pp. 1–20, 2023

    2023

  44. [52]

    Multi-scale weighted nuclear norm image restoration,

    N. Yair and T. Michaeli, “Multi-scale weighted nuclear norm image restoration,” inProceedings of the IEEE conference on computer vision and pattern recognition, 2018, pp. 3165–3174

    2018

  45. [53]

    Multi-stage progressive image restoration,

    S. W. Zamir, A. Arora, S. Khan, M. Hayat, F. S. Khan, M.-H. Yang, and L. Shao, “Multi-stage progressive image restoration,” inProceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2021, pp. 14821–14831

    2021

  46. [54]

    Multi-scale single image dehazing using laplacian and gaussian pyramids,

    Z. Li, H. Shu, and C. Zheng, “Multi-scale single image dehazing using laplacian and gaussian pyramids,”IEEE Transactions on Image Processing, vol. 30, pp. 9270–9279, 2021

    2021

  47. [55]

    Beyond gaussian pyramid: Multi-skip feature stacking for action recognition,

    Z. Lan, M. Lin, X. Li, A. G. Hauptmann, and B. Raj, “Beyond gaussian pyramid: Multi-skip feature stacking for action recognition,” in Proceedings of the IEEE conference on computer vision and pattern recognition, 2015, pp. 204–212

    2015

  48. [56]

    Iterative gaussian–laplacian pyramid network for hyperspectral image classification,

    C.-I. Chang, C.-C. Liang, and P. F. Hu, “Iterative gaussian–laplacian pyramid network for hyperspectral image classification,”IEEE Trans- actions on Geoscience and Remote Sensing, vol. 62, pp. 1–22, 2024

    2024

  49. [57]

    The laplacian pyramid as a compact image code,

    P. J. Burt and E. H. Adelson, “The laplacian pyramid as a compact image code,” in Readings in computer vision. Elsevier, 1987, pp. 671–679

    1987

  50. [58]

    Local laplacian filters: Edge- aware image processing with a laplacian pyramid

    S. Paris, S. W. Hasinoff, and J. Kautz, “Local laplacian filters: Edge- aware image processing with a laplacian pyramid.”ACM Trans. Graph., vol. 30, no. 4, p. 68, 2011

    2011

  51. [59]

    U-net: Convolutional net- works for biomedical image segmentation,

    O. Ronneberger, P. Fischer, and T. Brox, “U-net: Convolutional net- works for biomedical image segmentation,” inMedical image comput- ing and computer-assisted intervention–MICCAI 2015: 18th interna- tional conference, Munich, Germany, October 5-9, 2015, proceedings, part III ...

    2015

  52. [60]

    Ucl-dehaze: toward real-world image dehazing via unsupervised contrastive learning,

    Y. Wang, X. Yan, F. L. Wang, H. Xie, W. Yang, X.-P. Zhang, J. Qin, and M. Wei, “Ucl-dehaze: toward real-world image dehazing via unsupervised contrastive learning,” IEEE Transactions on Image Processing, vol. 33, pp. 1361–1374, 2024

    2024

  53. [61]

    Light-guided and cross-fusion u-net for anti-illumination image super-resolution,

    D. Cheng, L. Chen, C. Lv, L. Guo, and Q. Kou, “Light-guided and cross-fusion u-net for anti-illumination image super-resolution,”IEEE Transactions on Circuits and Systems for Video Technology, vol. 32, no. 12, pp. 8436–8449, 2022. JOURNAL OF LATEX CLASS FILES, VOL. 14, NO. 8, ...

    2022

  54. [62]

    Underwa- ter color restoration using u-net denoising autoencoder,

    Y. Hashisho, M. Albadawi, T. Krause, and U. F. von Lukas, “Underwa- ter color restoration using u-net denoising autoencoder,” in2019 11th International Symposium on Image and Signal Processing and Analysis (ISPA). IEEE, 2019, pp. 117–122

    2019

  55. [63]

    Feature pyramid networks for object detection,

    T.-Y. Lin, P. Dollár, R. Girshick, K. He, B. Hariharan, and S. Belongie, “Feature pyramid networks for object detection,” inProceedings of the IEEE conference on computer vision and pattern recognition, 2017, pp. 2117–2125

    2017

  56. [64]

    Parallel feature pyramid network for object detection,

    S.-W. Kim, H.-K. Kook, J.-Y. Sun, M.-C. Kang, and S.-J. Ko, “Parallel feature pyramid network for object detection,” inProceedings of the European conference on computer vision (ECCV), 2018, pp. 234–250

    2018

  57. [65]

    Pyramid attention network for image restoration,

    Y. Mei, Y. Fan, Y. Zhang, J. Yu, Y. Zhou, D. Liu, Y. Fu, T. S. Huang, and H. Shi, “Pyramid attention network for image restoration,” International Journal of Computer Vision, vol. 131, no. 12, pp. 3207– 3225, 2023

    2023

  58. [66]

    Uw- former: Underwater image enhancement via a semi-supervised multi- scale transformer,

    W. Chen, Y. Lei, S. Luo, Z. Zhou, M. Li, and C.-M. Pun, “Uw- former: Underwater image enhancement via a semi-supervised multi- scale transformer,” in2024 International Joint Conference on Neural Networks (IJCNN). IEEE, 2024, pp. 1–8

    2024

  59. [67]

    Sguie-net: Semantic attention guided underwater image enhancement with multi- scale perception,

    Q. Qi, K. Li, H. Zheng, X. Gao, G. Hou, and K. Sun, “Sguie-net: Semantic attention guided underwater image enhancement with multi- scale perception,”IEEE Transactions on Image Processing, vol. 31, pp. 6816–6830, 2022

    2022

  60. [68]

    Hierarchical attention aggregation with multi-resolution feature learning for gan- based underwater image enhancement,

    D. Zhang, C. Wu, J. Zhou, W. Zhang, C. Li, and Z. Lin, “Hierarchical attention aggregation with multi-resolution feature learning for gan- based underwater image enhancement,” Engineering Applications of Artificial Intelligence, vol. 125, p. 106743, 2023

    2023

  61. [69]

    Mffn: An underwater sensing scene im- age enhancement method based on multiscale feature fusion network,

    R. Chen, Z. Cai, and W. Cao, “Mffn: An underwater sensing scene im- age enhancement method based on multiscale feature fusion network,” IEEE Transactions on Geoscience and Remote Sensing, vol. 60, pp. 1–12, 2021

    2021

  62. [70]

    Underwater image enhance- ment via extended multi-scale retinex,

    S. Zhang, T. Wang, J. Dong, and H. Yu, “Underwater image enhance- ment via extended multi-scale retinex,”Neurocomputing, vol. 245, pp. 1–9, 2017

    2017

  63. [71]

    Domain adaptation for underwater image enhancement,

    Z. Wang, L. Shen, M. Xu, M. Yu, K. Wang, and Y. Lin, “Domain adaptation for underwater image enhancement,”IEEE Transactions on Image Processing, vol. 32, pp. 1442–1457, 2023

    2023

  64. [72]

    Edge-computing-enabled deep learning approach for low-light satellite image enhancement,

    T.-A. Bui, P.-J. Lee, C.-S. Liang, P.-H. Hsu, S.-H. Shiu, and C.-K. Tsai, “Edge-computing-enabled deep learning approach for low-light satellite image enhancement,”IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, vol. 17, pp. 4071–4083, 2024

    2024

  65. [73]

    Remote sens- ing image super-resolution with residual split attention mechanism,

    X. Chen, Y. Wu, T. Lu, Q. Kong, J. Wang, and Y. Wang, “Remote sens- ing image super-resolution with residual split attention mechanism,” IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, vol. 16, pp. 1–13, 2023

    2023

  66. [74]

    Act-sr:Aggregation connection transformer for remote sensing image super-resolution,

    Y.Kang,X. Wang,X.Zhang,S. Wang,andG.Jin, “Act-sr:Aggregation connection transformer for remote sensing image super-resolution,” IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2024

    2024

  67. [75]

    Hyda- net: A hybrid dense attention network for remote sensing multi-image super-resolution,

    M. R. Ibrahim, R. Benavente, D. Ponsa, and F. Lumbreras, “Hyda- net: A hybrid dense attention network for remote sensing multi-image super-resolution,” IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2025

    2025

  68. [76]

    Atwo-branchmulti- scale residual attention network for single image super-resolution in remote sensing imagery,

    A.Patnaik,M.K.Bhuyan,andK.F.MacDorman,“Atwo-branchmulti- scale residual attention network for single image super-resolution in remote sensing imagery,”IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2024

    2024

  69. [77]

    Spectral– spatial generative adversarial network for super-resolution land cover mapping with multispectral remotely sensed imagery,

    C. Shang, S. Jiang, F. Ling, X. Li, Y. Zhou, and Y. Du, “Spectral– spatial generative adversarial network for super-resolution land cover mapping with multispectral remotely sensed imagery,”IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, vol. ...

    2022

  70. [78]

    A rapid scene depth estimation model based on underwater light attenuation prior for underwater image restoration,

    W. Song, Y. Wang, D. Huang, and D. Tjondronegoro, “A rapid scene depth estimation model based on underwater light attenuation prior for underwater image restoration,” in Advances in Multimedia Information Processing–PCM 2018: 19th Pacific-Rim Conference on Multimedia, Hefei, C...

    2018

  71. [79]

    Springer, 2018, pp. 678–688

    2018

  72. [80]

    Enhancing underwater images and videos by fusion,

    C. Ancuti, C. O. Ancuti, T. Haber, and P. Bekaert, “Enhancing underwater images and videos by fusion,” in2012 IEEE conference on computer vision and pattern recognition. IEEE, 2012, pp. 81–88

    2012

  73. [81]

    Pdr-net: Perception-inspired single image dehazing network with refinement,

    C. Li, C. Guo, J. Guo, P. Han, H. Fu, and R. Cong, “Pdr-net: Perception-inspired single image dehazing network with refinement,” IEEE Transactions on Multimedia, vol. 22, no. 3, pp. 704–716, 2020

    2020

  74. [82]

    Contrastive semi- supervised learning for underwater image restoration via reliable bank,

    S. Huang, K. Wang, H. Liu, J. Chen, and Y. Li, “Contrastive semi- supervised learning for underwater image restoration via reliable bank,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2023, pp. 18145–18155

    2023

  75. [83]

    Diffwater: Underwater image enhancement based on conditional de- noising diffusion probabilistic model,

    M. Guan, H. Xu, G. Jiang, M. Yu, Y. Chen, T. Luo, and X. Zhang, “Diffwater: Underwater image enhancement based on conditional de- noising diffusion probabilistic model,”IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, vol. 17, pp. 2319– 2335, 2023

    2023

  76. [84]

    Ugif-net: An Efficient Fully Guided Information Flow Network for Underwater Image Enhancement,

    J. Zhou, B. Li, D. Zhang, J. Yuan, W. Zhang, Z. Cai, and J. Shi, “Ugif-net: An Efficient Fully Guided Information Flow Network for Underwater Image Enhancement,”IEEE Transactions on Geoscience and Remote Sensing, 2023

    2023

  77. [85]

    U-Net: Convolutional Networks for Biomedical Image Segmentation,

    O. Ronneberger, P. Fischer, and T. Brox, “U-Net: Convolutional Networks for Biomedical Image Segmentation,” in Medical Image Computing and Computer-Assisted Intervention. Springer, 2015, pp. 234–241

    2015

  78. [86]

    Highway networks,

    R. K. Srivastava, K. Greff, and J. Schmidhuber, “Highway networks,” arXiv preprint arXiv:1505.00387, 2015

    2015 arXiv

  79. [87]

    Deep residual learning for image recognition,

    K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for image recognition,”in ProceedingsoftheIEEEconferenceoncomputervision and pattern recognition, 2016, pp. 770–778

    2016

  80. [88]

    Multi-residual networks: Improv- ing the speed and accuracy of residual networks,

    M. Abdi and S. Nahavandi, “Multi-residual networks: Improv- ing the speed and accuracy of residual networks,” arXiv preprint arXiv:1609.05672, 2016

    2016 arXiv

  81. [89]

    Selective kernel networks,

    X. Li, W. Wang, X. Hu, and J. Yang, “Selective kernel networks,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2019, pp. 510–519

    2019

  82. [90]

    Recurrent models of visualattention,

    V. Mnih, N. Heess, A. Graves, and K. Kavukcuoglu, “Recurrent models of visualattention,”Advances inneural informationprocessing systems, vol. 27, 2014

    2014

  83. [91]

    Neural machine translation by jointly learning to align and translate,

    D. Bahdanau, K. Cho, and Y. Bengio, “Neural machine translation by jointly learning to align and translate,”arXiv preprint arXiv:1409.0473, 2014

    2014 arXiv

  84. [92]

    Show, attend and tell: Neural image caption generation with visual attention,

    K. Xu, J. Ba, R. Kiros, K. Cho, A. Courville, R. Salakhudinov, R. Zemel, and Y. Bengio, “Show, attend and tell: Neural image caption generation with visual attention,” inInternational conference on machine learning. PMLR, 2015, pp. 2048–2057

    2015

  85. [93]

    Attention is All You Need,

    A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, Ł. Kaiser, and I. Polosukhin, “Attention is All You Need,” Advances in Neural Information Processing Systems, vol. 30, 2017

    2017

  86. [94]

    Generative adversarial nets,

    I. J. Goodfellow, J. Pouget-Abadie, M. Mirza, B. Xu, D. Warde-Farley, S. Ozair, A. Courville, and Y. Bengio, “Generative adversarial nets,” Advances in neural information processing systems, vol. 27, 2014

    2014

  87. [95]

    Tcrn: A two-step underwater image enhancement network based on triple-color space feature recon- struction,

    S. Lin, R. Zhang, Z. Ning, and J. Luo, “Tcrn: A two-step underwater image enhancement network based on triple-color space feature recon- struction,” Journal of Marine Science and Engineering, vol. 11, no. 6, p. 1221, 2023

    2023

  88. [96]

    Coc-ufgan: Underwater image enhancement based on color opponent compensation and dual-subnet underwater fusion generative adversarial network,

    Z. Liu, X. Fu, C. Lin, and H. Xu, “Coc-ufgan: Underwater image enhancement based on color opponent compensation and dual-subnet underwater fusion generative adversarial network,”Journal of Visual Communication and Image Representation, vol. 100, p. 104101, 2024

    2024

  89. [97]

    Nas-fpn: Learning scalable feature pyramid architecture for object detection,

    G. Ghiasi, T.-Y. Lin, and Q. V. Le, “Nas-fpn: Learning scalable feature pyramid architecture for object detection,” inProceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2019, pp. 7036–7045

    2019

  90. [98]

    Five A + Network: You Only Need 9K Parameters for Underwater Image Enhancement,

    J. Jiang, T. Ye, J. Bai, S. Chen, W. Chai, S. Jun, Y. Liu, and E. Chen, “Five A + Network: You Only Need 9K Parameters for Underwater Image Enhancement,”arXiv preprint arXiv:2305.08824, 2023

    2023 arXiv

  91. [99]

    A fusion adversarial underwater im- age enhancement network with a public test dataset,

    H. Li, J. Li, and W. Wang, “A fusion adversarial underwater im- age enhancement network with a public test dataset,”arXiv preprint arXiv:1906.06819, 2019

    1906 arXiv

  92. [100]

    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 Technol- ogy, vol. 30, no. 12, pp. 4861–4875, 2020

    2020

  93. [101]

    The unreasonable effectiveness of deep features as a perceptual metric,

    R. Zhang, P. Isola, A. A. Efros, E. Shechtman, and O. Wang, “The unreasonable effectiveness of deep features as a perceptual metric,” in Proceedings of the IEEE conference on computer vision and pattern recognition, 2018, pp. 586–595. JOURNAL OF LATEX CLASS FILES, VOL. 14, NO....

    2018

  94. [2019]

    Yakun Ju is currently an Assistant Professor (UK Lecturer) with the School of Computing and Math- ematical Sciences, University of Leicester, U.K

    His current research interests include compu- tational intelligence, image processing, system iden- tification, and nonlinear time-series analysis. Yakun Ju is currently an Assistant Professor (UK Lecturer) with the School of Computing and Math- ematical Sciences, University o...

    2016

  95. [2024]

    Currently, his research focuses on several cutting-edge areas, including variational image pro- cessing, medical image processing, and convex op- timization

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.