Pith. sign in

REVIEW 4 major objections 4 minor 59 references

Human Vision Constrained Super-Resolution

T0 review · 4 major / 4 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read Cut super-resolution compute by skipping detail the eye can't see

desk verdict A genuinely HVS-grounded approach to adaptive SR with promising user evidence, but the core derivation has a sign error that needs fixing before the results reproduce. read the letter →

arxiv 2411.17513 v3 pith:GARNGGR3 submitted 2024-11-26 cs.CV cs.GRcs.LG

Volodymyr Karpenko , Taimoor Tariq , Jorge Condor , Piotr Didyk This is my paper · ORCID
classification cs.CVcs.GRcs.LG
keywords humanvisualsystemsuper-resolutioncontrastsensitivityfunctionmaskingjust-noticeabledifferencenetworkbranchingcomputationalefficiencyfoveatedrendering
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper sets out to show that neural super-resolution spends most of its compute on detail the human eye cannot resolve, and that this waste can be eliminated without visible quality loss. It builds a Human Visual Processing Framework (HVPF) that, for each image patch, computes the strongest frequency attenuation a viewer would not notice, then routes the patch to the lightest network branch or shallowest network whose measured frequency response stays within that bound. On VDSR and EDSR, the method cuts FLOPS by factors of two to four for both 2x and 4x upsampling. Two-alternative forced-choice user studies found that observers could not tell the adaptive output from the full-network output, with preference hovering near 50 percent. The paper also extends the same control signal to video motion and to gaze-dependent peripheral vision for VR/AR displays.

What carries the argument

The load-bearing device is the Human Visual Processing Framework (HVPF), which functions as a visual-difference predictor turned into a compute allocator. Its inputs are the precomputed Gaussian-fitted attenuation curves of each SR variant and, for each patch, a three-level luminance-contrast pyramid normalized by the contrast sensitivity function and weighted by a visual masking term. The core identity is the closed-form solution of the 1 JND constraint, Eq. (7), which expresses the tolerable attenuated contrast $C'_n(f,p)$ directly in terms of the input patch's masked contrast $C_n$, thanks to the assumption that masking does not change under upsampling. The selection step then chooses the variant whose attenuation vector has the highest cosine similarity to the tolerable attenuation vector of the patch. The attenuation curve itself links network capacity to spatial-frequency response, which is what makes the framework independent of any specific network architecture.

What would settle it

Run a forced-choice detection study on textured patches at the model's predicted 1 JND attenuation: if observers reliably discriminate the attenuated output from the full-network output on patches where the network adds substantial high-frequency energy, the masking approximation is failing. A sharper test is to compare the model's predicted tolerable attenuation against measured detection thresholds on synthetic patches with controlled masking, where the true threshold can be established psychophysically.

Watch

Extended reading notes

Core claim

The central claim is that a super-resolution network can be run at a fraction of its compute with no perceivable drop in quality, provided the compute is allocated per image patch according to what the human visual system can resolve. The paper measures each candidate SR variant by an attenuation curve, defined as the radially averaged ratio of the Fourier magnitude of the network output to that of the ground truth, and fits these curves with a Gaussian fall-off. For each input patch, the HVPF builds a Laplacian pyramid of luminance contrast, normalizes it by a contrast sensitivity function, applies a visual masking model, and solves the constraint that the perceived contrast change be exactly 1 JND, under the approximation that masking on the upsampled output equals masking on the input. This yields a tolerable attenuation vector, and the branch or network with the most similar attenuation curve is selected. In their experiments with branched VDSR and channel-scaled EDSR, this selection reduced FLOPS by 51 to 78 percent while user-study preference stayed near chance; the authors claim the framework is architecture-agnostic and extends to video and foveated displays.

Load-bearing premise

The load-bearing premise is that the visual masking measured on the low-resolution input patch is a faithful stand-in for the masking of the network's upsampled output, so the tolerable attenuation can be computed from the input alone; if upsampling changes the local masking enough, the 1 JND bound is miscalibrated and the branch choice goes wrong.

Editorial extensions

If this is right

  • The same 1 JND control signal can steer other efficiency knobs in SR, such as per-patch quantization, pruning, or input resolution, not just branching and depth.
  • SR networks can be designed with early-exit branches from the start, because the framework tells each branch which patches it can serve without visible loss.
  • In video, the framework spends less spatial compute on fast-moving regions, where temporal frequency already reduces acuity, on top of patch-level savings.
  • For VR/AR displays, the eccentricity-aware CSF enables gaze-contingent SR that spends full quality at the fovea and much less in the periphery.
  • If adopted broadly, the criterion for SR quality shifts from pixel error to perceptual cost, making FLOPS-per-JND a natural efficiency metric.

Reading between the lines

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

  • One extension the authors do not run: applying HVPF to generative or diffusion SR, whose synthesized high frequencies are invented rather than preserved; the attenuation proxy measures signal presence, not correctness, so the 1 JND calibration would need a correctness-aware term.
  • Because Eq. (7) is closed-form, the whole branch-selection pipeline is differentiable; one could end-to-end train an SR family with the 1 JND constraint as the loss, potentially tightening the compute/quality trade-off beyond the hand-fitted curves.
  • The framework suggests a shared perceptual budget across a video codec and a decoder-side SR network, coupling bit allocation and upsampling compute to the same CSF-based target.
  • A stress test worth running is content outside natural-image statistics (text, faces, medical imagery), where the average attenuation-curve prior may not match the actual network response and the 1 JND bound could be off.
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 / 4 minor

Summary. The paper proposes a Human Visual Processing Framework (HVPF) that dynamically selects, per image patch, how much of an SR network's computation is needed before the output reaches a 1-JND perceptual limit. The method precomputes frequency-dependent attenuation curves for SR variants (early-exit branches of VDSR, or EDSR variants of different depths), fits them with Gaussians, and then uses a CSF-plus-masking contrast model to compute a tolerable attenuation vector for each input patch. Branch/network selection is performed by matching this vector to the precomputed curves. The authors report PSNR/SSIM, FLOPS reductions of 2x and greater, 2AFC user studies on images and video, and preliminary results for foveated AR/VR SR.

Significance. If the claims hold, the paper offers a practical, architecture-agnostic way to reduce SR inference cost without perceptible quality loss. The strengths are the explicit use of a calibrated low-level vision model (CSF, contrast masking, visual difference prediction) as a per-patch policy, the validation on two classic SR architectures (VDSR and EDSR), the inclusion of video and eccentricity-aware extensions, and the use of human-subject experiments rather than only image-quality metrics. The reported FLOPS reductions in Table 1 are substantial and, if reproducible end-to-end, would be useful for real-time SR applications. However, the central derivation of the tolerable attenuation contains a sign inconsistency that propagates into the branch-selection rule, the subjective equivalence claim is not backed by inferential statistics, and the reported FLOPS figures exclude the HVPF overhead. These points need to be resolved before the main claim of 'no perceived quality loss at 2x or greater FLOPS reduction' can be taken as established.

major comments (4)
  1. [Sec. 4.3, Eq. (4)-(7)] The derivation of the tolerable attenuation is algebraically inconsistent. Eq. (4) is written as C'_t(p,f) - C_t(p,f) = 1, which states that the output has one JND more, not less, perceived contrast than the input. Under the masking-equality assumption, the literal reading of Eq. (4) gives |C'_n|^alpha = |C_n|^alpha + M with M = 1 + (1/|N|) sum |C_n(q)|^beta, so every t_i in Eq. (8) would exceed 1. The printed Eq. (7), |M - |C_n|^alpha|^{1/alpha}, corresponds instead to the attenuation case |C'_n|^alpha = |C_n|^alpha - M, and only when |C_n|^alpha >= M. The text never states that Eq. (4) is meant to be an absolute difference, nor what should be done when no attenuation root exists. Since Eq. (8) computes t_i from C'_n, and Eq. (9) uses t_i for the final branch/network selection, this ambiguity is load-bearing: the reported FLOPS savings and user-study outcomes cannot be traced to the published algorithm until the sign is corrected and the one-sided versus two-sided nature of the JND constraint is resolved.
  2. [Sec. 6.2, Figs. 5, 7, 8] The central subjective claim is supported only by mean preference proportions near 50% (e.g., 49.07% and 52.04% in Fig. 7 and Fig. 8), with no confidence intervals, no per-observer variance, and no statistical test. A 2AFC proportion near 50% does not establish perceptual equivalence; it is also compatible with a study that is underpowered to detect a real difference. Because the paper explicitly argues that 'the results indeed conform with our hypothesis that there is no perceivable loss in visual quality,' the authors should report binomial or bootstrap confidence intervals, per-participant distributions, and preferably an equivalence test (e.g., TOST with a pre-specified margin) for the preference proportions. Without this, the user studies remain suggestive but not statistically decisive.
  3. [Sec. 4.4, Tables 1-2] The reported FLOPS reductions appear to be network-only: the cost of the HVPF itself (contrast pyramid construction, Eq. (7), and the matching in Eq. (9)) is not included in any of the percentages in Tables 1 and 2. Section 4.4 defends the feasibility of a GPU implementation by citing prior work, but the abstract and conclusions make quantitative claims about 'reducing FLOPS' that should be verified end-to-end. Please either include the HVPF FLOPS in the reported totals, or clearly state that the percentages are network-only and provide an upper-bound estimate of the HVPF overhead. Without this, the magnitude of the claimed efficiency gain is not fully quantified.
  4. [Sec. 4.3, Eq. (9)] The branch-selection rule in Eq. (9) uses cosine similarity between the tolerable attenuation vector t and each candidate attenuation vector t_hat_j. Cosine similarity is scale-invariant, so a branch whose attenuation curve has the same shape as t but is everywhere more attenuating receives the same score as a branch with the same shape and less attenuation. The optimization therefore does not directly enforce the 1-JND constraint, nor does it minimize computation subject to that constraint. Please justify the cosine similarity choice, replace it with a scale-sensitive criterion (e.g., L2 distance), or add a constraint that the selected branch's attenuation does not exceed the tolerable attenuation at any frequency.
minor comments (4)
  1. [Sec. 4.1, Eq. (3)] The Gaussian fits to the attenuation curves are used as the selection templates in Eq. (9), but no fit quality or residual error is reported. Please report the fitting error or show the fitted curves versus the measured curves.
  2. [Sec. 4.3, Eq. (8)] The paper uses three levels of the contrast pyramid without justifying this choice. Please explain how t_1, t_2, t_3 correspond to the three spatial frequencies and whether the result is sensitive to the number of pyramid levels.
  3. [Sec. 5 and Supplementary Sec. D] The per-patch computation of HVPF is not fully specified at patch boundaries. If the Laplacian pyramid is computed on the whole low-resolution image, please describe the padding or overlap strategy; if it is computed independently per patch, the contrast-masking neighborhood in Eq. (6) can cross patch borders inconsistently. Please clarify.
  4. [Secs. 4.1 and Table 1] The attenuation curves are fitted on Set5 and Set14, and the same datasets appear in Table 1. Although the curves are method-level characteristics, evaluating on the calibration set is a mild form of tuning. Please report whether the conclusions change when the curves are fitted on one dataset and evaluated on another, or explicitly state that the curves are calibration-free after fitting.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity: the branch-selection rule matches an independently derived JND-based attenuation target to precomputed network attenuation curves, and the central claim is validated by external user studies.

full rationale

The paper's derivation chain is not circular. The attenuation curves in Eqs. (1)-(3) are measured properties of the SR variants, computed once on a set of natural images and fitted with Gaussians. The tolerable attenuation t in Eq. (8) is derived from the VDP contrast model in Eq. (6), with constants taken from external vision literature (alpha=0.7, beta=0.2, CSF from StelaCSF, masking from Zeng et al.). The branch/network selection in Eq. (9) is a nearest-neighbor match between this independently computed target t and the precomputed attenuation curves. There is no fitted parameter that is renamed as a prediction: t does not depend on the fitted curves, and the curves are not fitted to the user-study outcomes. The central claim of no perceived quality loss is tested by 2AFC user studies against the full-network output, which is an external benchmark independent of the predictor's fitted values. The self-citations present ([46] Tursun et al., which includes a present author, for the contrast-masking approximation; [45] Tariq et al., which includes a present author, for GPU overhead) are load-bearing in parts of the derivation and efficiency argument, but both are published external works and the masking approximation is explicitly stated and validated end-to-end by the user studies. These are self-citations but not circularity. One correctness risk is noted but does not affect circularity: Eq. (4) states C'_t - C_t = 1, while Eq. (7) solves the opposite attenuation direction, C_t - C'_t = 1 (or an absolute-value reading); this sign ambiguity is a derivation flaw that could miscalibrate branch selection, but it is not an equivalence between inputs and outputs. Overall, the method is self-contained against external human judgment, so the circularity score is low.

Assumptions & free parameters 2 free parameters · 7 assumptions · 0 invented entities

The central model rests on assumptions from prior vision science (CSF, masking, 1 JND threshold) plus a fitted Gaussian representation of network attenuation and the assertion that network Fourier attenuation transfers across images. No new physical entities are introduced. The main unmeasured cost is the HVPF itself, whose FLOP overhead is not included in the reported efficiency numbers.

free parameters (2)
  • Gaussian attenuation parameters (a, b, c) per SR variant and scale = not reported
    Fitted in Eq (3) from attenuation curves over 19 natural images in Section 4.1; used in Eq (9) for branch selection.
  • Contrast masking exponents alpha and beta = alpha=0.7, beta=0.2
    Fixed in Section 4.2 from Zeng et al. [55]; part of the JND model and not fitted in this paper.
assumptions (7)
  • domain assumption Average attenuation curves computed on 19 natural images transfer to arbitrary input patches.
    Section 4.1 states the average spectrum of natural images follows an inverse power law and therefore an average attenuation curve is a good approximation; this is used to select branches for unseen patches.
  • domain assumption Fourier magnitude ratio (Eq 1) is a sufficient proxy for reconstruction quality.
    Section 4.1 acknowledges the measure quantifies signal presence, not correctness, but uses it as a quality proxy; Eq (5) then treats attenuation as a modulation of physical contrast.
  • domain assumption Visual masking of the upsampled output equals visual masking of the input patch.
    Section 4.3, immediately before Eq (7): 'we make the assumption that the contrast masking for the up-sampled output patch can be approximated by that of the input patch'; this is necessary to solve Eq (6) in closed form.
  • domain assumption A 1 JND difference is the correct perceptual threshold for setting the tolerable attenuation.
    Section 4.3 sets Eq (4) to exactly 1 JND; this assumes the CSF and masking model predict SR artifact visibility accurately at this threshold.
  • ad hoc to paper The Gaussian fall-off model (Eq 3) accurately represents attenuation curves.
    The parameters a, b, c are fitted per SR variant and used in Eq (9) to select branches; no goodness-of-fit is reported.
  • domain assumption Per-patch independent processing does not create visible spatial or temporal inconsistencies.
    Section 7 admits consistency is not guaranteed and aliasing was observed; the authors rely on participants not reporting issues instead of a measured guarantee.
  • domain assumption Temporal sensitivity can be captured by converting optical flow to a temporal frequency at 24 fps.
    Section 6.1 estimates temporal frequency from optical flow at 24 fps and uses it to modulate the CSF; no validation of this conversion is provided.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Human Vision Constrained Super-Resolution." pith.science (2026). https://pith.science/paper/GARNGGR3

@misc{pith2026241117513,
  author       = {Pith},
  title        = {Pith review of: Human Vision Constrained Super-Resolution},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/GARNGGR3}},
  note         = {Machine review of arXiv:2411.17513}
}
abstract

Modern deep-learning super-resolution (SR) techniques process images and videos independently of the underlying content and viewing conditions. However, the sensitivity of the human visual system (HVS) to image details changes depending on the underlying image characteristics, such as spatial frequency, luminance, color, contrast, or motion; as well viewing condition aspects such as ambient lighting and distance to the display. This observation suggests that computational resources spent on up-sampling images/videos may be wasted whenever a viewer cannot resolve the synthesized details i.e the resolution of details exceeds the resolving capability of human vision. Motivated by this observation, we propose a human vision inspired and architecture-agnostic approach for controlling SR techniques to deliver visually optimal results while limiting computational complexity. Its core is an explicit Human Visual Processing Framework (HVPF) that dynamically and locally guides SR methods according to human sensitivity to specific image details and viewing conditions. We demonstrate the application of our framework in combination with network branching to improve the computational efficiency of SR methods. Quantitative and qualitative evaluations, including user studies, demonstrate the effectiveness of our approach in reducing FLOPS by factors of 2$\times$ and greater, without sacrificing perceived quality.

Figures

Figures reproduced from arXiv: 2411.17513 by the authors.

Figure 1
Figure 1. Visual results of our method compared to the original networks. On the right, we can observe the maps produced by our [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. The attenuation curves derived for the case of modulat [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 4
Figure 4. Attenuation curves derived from the SwinIR, a trans [PITH_FULL_IMAGE:figures/full_fig_p006_4.png] view at source ↗
Figures from the paper (12 more)
Figure 3
Figure 3. Figure 3: Flowchart illustrating the methodology employed for ef [PITH_FULL_IMAGE:figures/full_fig_p006_3.png]
Figure 5
Figure 5. Figure 5: The result of our subjective study (for 15 participants) [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 6
Figure 6. Figure 6: Our model predictions based on gaze position with [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]
Figure 7
Figure 7. Figure 7: The result of our subjective study (for 9 participants) for [PITH_FULL_IMAGE:figures/full_fig_p012_7.png]
Figure 8
Figure 8. Figure 8: The result of our subjective study (for 14 participants) [PITH_FULL_IMAGE:figures/full_fig_p012_8.png]
Figure 9
Figure 9. Figure 9: Our model predictions based on gaze position with [PITH_FULL_IMAGE:figures/full_fig_p014_9.png]
Figure 10
Figure 10. Figure 10: Our method quality map prediction results for [PITH_FULL_IMAGE:figures/full_fig_p015_10.png]
Figure 11
Figure 11. Figure 11: Visual results of our method compared to the original networks. On the right, we can observe the maps produced by our [PITH_FULL_IMAGE:figures/full_fig_p016_11.png]
Figure 12
Figure 12. Figure 12: The SR result (×4) of VDSR network and our method applied to VDSR network for image DIV2K-0885. 5 [PITH_FULL_IMAGE:figures/full_fig_p016_12.png]
Figure 13
Figure 13. Figure 13: The SR result (×4) of VDSR network and our method applied to VDSR network for image DIV2K-0878. (a) Original Network (b) Our method [PITH_FULL_IMAGE:figures/full_fig_p017_13.png]
Figure 14
Figure 14. Figure 14: The SR result (×4) of VDSR network and our method applied to VDSR network for image DIV2K-0850. 6 [PITH_FULL_IMAGE:figures/full_fig_p017_14.png]
Figure 15
Figure 15. Figure 15: The SR result (×4) of VDSR network and our method applied to VDSR network for image DIV2K-0815. 7 [PITH_FULL_IMAGE:figures/full_fig_p018_15.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

59 extracted references · 52 canonical work pages

  1. [1]

    Lightroom super resolution

    Adobe Inc. Lightroom super resolution. https://www. adobe.com/products/photoshop- lightroom/ 8 super-resolution.html, 2024. Accessed: 2024-11-

    work page 2024

  2. [2]

    Fidelityfx super resolution (fsr)

    AMD Corporation. Fidelityfx super resolution (fsr). https : / / www . amd . com / en / technologies / radeon - software - fidelityfx - super - resolution, 2024. Accessed: 2024-11-15. 1, 3

    work page 2024

  3. [3]

    Mantiuk, Alexandre Chapiro, and Sophie Wuerger

    Maliha Ashraf, Rafał K. Mantiuk, Alexandre Chapiro, and Sophie Wuerger. castlecsf — a contrast sensitivity function of color, area, spatiotemporal frequency, luminance and ec- centricity. Journal of Vision, 24(4):5–5, 2024. 8

    work page 2024

  4. [4]

    Mantiuk, Alexandre Chapiro, and Sophie Wuerger

    Maliha Ashraf, Rafał K. Mantiuk, Alexandre Chapiro, and Sophie Wuerger. castlecsf — a contrast sensitivity function of color, area, spatiotemporal frequency, luminance and ec- centricity. Journal of Vision, 24, 2024. 3

    work page 2024

  5. [5]

    SRTransGAN: Image Super-Resolution using Transformer based Generative Adversarial Network

    Neeraj Baghel, Shiv Ram Dubey, and Satish Kumar Singh. Srtransgan: Image super-resolution using transformer based generative adversarial network. ArXiv, abs/2312.01999,

    work page Pith review arXiv

  6. [6]

    Peter G. J. Barten. Formula for the contrast sensitivity of the human eye. In IS&T/SPIE Electronic Imaging, 2003. 3

    work page 2003

  7. [7]

    Guillemot, and Marie-Line Alberi-Morel

    Marco Bevilacqua, Aline Roumy, Christine M. Guillemot, and Marie-Line Alberi-Morel. Low-complexity single- image super-resolution based on nonnegative neighbor em- bedding. In British Machine Vision Conference, 2012. 5

    work page 2012

  8. [8]

    Bicubic++: Slim, slimmer, slimmest-designing an industry-grade super- resolution network

    Bahri Batuhan Bilecen and Mustafa Ayazoglu. Bicubic++: Slim, slimmer, slimmest-designing an industry-grade super- resolution network. In Proceedings of the IEEE/CVF Con- ference on Computer Vision and Pattern Recognition, pages 1623–1632, 2023. 1

    work page 2023

Show all 59 references
  1. [9]

    Burt and E

    P. Burt and E. Adelson. The laplacian pyramid as a compact image code. IEEE Transactions on Communications, 31(4): 532–540, 1983. 5

    1983

  2. [10]

    Topography of ganglion cells in human retina

    C A Curcio and K A Allen. Topography of ganglion cells in human retina. J Comp Neurol, 300(1):5–25, 1990. 2

    1990

  3. [11]

    Scott J. Daly. Visible differences predictor: an algorithm for the assessment of image fidelity. In Electronic imaging,

  4. [12]

    Imagenet: A large-scale hierarchical image database

    Jia Deng, Wei Dong, Richard Socher, Li-Jia Li, Kai Li, and Li Fei-Fei. Imagenet: A large-scale hierarchical image database. In 2009 IEEE Conference on Computer Vision and Pattern Recognition, pages 248–255, 2009. 3

    2009

  5. [13]

    Image super-resolution using deep convolutional net- works

    Chao Dong, Chen Change Loy, Kaiming He, and Xiaoou Tang. Image super-resolution using deep convolutional net- works. IEEE Transactions on Pattern Analysis and Machine Intelligence, 38:295–307, 2014. 3

    2014

  6. [14]

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

    Alexey Dosovitskiy, Lucas Beyer, Alexander Kolesnikov, Dirk Weissenborn, Xiaohua Zhai, Thomas Unterthiner, Mostafa Dehghani, Matthias Minderer, Georg Heigold, Syl- vain Gelly, Jakob Uszkoreit, and Neil Houlsby. An image is worth 16x16 words: Transformers for image recognition ...

    2010 arXiv

  7. [15]

    Generative adversarial networks

    Ian Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron Courville, and Yoshua Bengio. Generative adversarial networks. Commun. ACM, 63(11):139–144, 2020. 3

    2020

  8. [16]

    Comparison of single image hdr reconstruction methods — the caveats of quality assessment

    Param Hanji, Rafal Mantiuk, Gabriel Eilertsen, Saghi Ha- jisharif, and Jonas Unger. Comparison of single image hdr reconstruction methods — the caveats of quality assessment. In ACM SIGGRAPH 2022 Conference Proceedings , New York, NY , USA, 2022. Association for Computing Mach...

    2022

  9. [17]

    Zhang, Shaoqing Ren, and Jian Sun

    Kaiming He, X. Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition. 2016 IEEE Confer- ence on Computer Vision and Pattern Recognition (CVPR) , pages 770–778, 2015. 3

    2016

  10. [18]

    Xe super sampling (xess)

    Intel Corporation. Xe super sampling (xess). https : / / www . intel . com / content / www / us / en / products / docs / discrete - gpus / arc / technology/xess.html , 2024. Accessed: 2024-11-

    2024

  11. [19]

    Accelerating image super-resolution networks with pixel- level classification

    Jinho Jeong, Jinwoo Kim, Younghyun Jo, and Seon Joo Kim. Accelerating image super-resolution networks with pixel- level classification. In European Conference on Computer Vision, pages 236–251. Springer, 2024. 2

    2024

  12. [20]

    Depth-controllable very deep super-resolution network

    Dohyun Kim, Joongheon Kim, Junseok Kwon, and Tae- Hyung Kim. Depth-controllable very deep super-resolution network. In 2019 International Joint Conference on Neural Networks (IJCNN), pages 1–8, 2019. 6

    2019

  13. [21]

    Accurate image super-resolution using very deep convolutional net- works

    Jiwon Kim, Jung Lee, and Kyoung Mu Lee. Accurate image super-resolution using very deep convolutional net- works. 2015. 3

    2015

  14. [22]

    Accurate image super-resolution using very deep convolutional net- works

    Jiwon Kim, Jung Kwon Lee, and Kyoung Mu Lee. Accurate image super-resolution using very deep convolutional net- works. IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pages 1646–1654, 2016. 6

    2016

  15. [23]

    Classsr: A general framework to accelerate super- resolution networks by data characteristic

    Xiangtao Kong, Hengyuan Zhao, Yu Qiao, and Chao Dong. Classsr: A general framework to accelerate super- resolution networks by data characteristic. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 12016–12025, 2021. 2

    2021

  16. [24]

    Alex Krizhevsky, Ilya Sutskever, and Geoffrey E. Hinton. Imagenet classification with deep convolutional neural net- works. In Proceedings of the 25th International Conference on Neural Information Processing Systems - Volume 1, page 1097–1105, Red Hook, NY , USA, 2012. Curran...

    2012

  17. [25]

    Towards high-quality and efficient video super-resolution via spatial-temporal data overfitting

    Gen Li, Jie Ji, Minghai Qin, Wei Niu, Bin Ren, Fatemeh Afghah, Lin Guo, and Xiaolong Ma. Towards high-quality and efficient video super-resolution via spatial-temporal data overfitting. 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 10259–102...

    2023

  18. [26]

    Srdiff: Single image super- resolution with diffusion probabilistic models

    Haoying Li, Yifan Yang, Meng Chang, Huajun Feng, Zhi hai Xu, Qi Li, and Yue ting Chen. Srdiff: Single image super- resolution with diffusion probabilistic models. Neurocom- puting, 479:47–59, 2021. 3

    2021

  19. [27]

    Zhang, Luc Van Gool, and Radu Timofte

    Jingyun Liang, Jie Cao, Guolei Sun, K. Zhang, Luc Van Gool, and Radu Timofte. Swinir: Image restoration using swin transformer. 2021 IEEE/CVF International Conference on Computer Vision Workshops (ICCVW), pages 1833–1844,

    2021

  20. [28]

    Enhanced deep residual networks for single image super-resolution, 2017

    Bee Lim, Sanghyun Son, Heewon Kim, Seungjun Nah, and Kyoung Mu Lee. Enhanced deep residual networks for single image super-resolution, 2017. 6 9

    2017

  21. [29]

    Transformer for single image super-resolution

    Zhisheng Lu, Juncheng Li, Hong Liu, Chao Huang, Linlin Zhang, and Tieyong Zeng. Transformer for single image super-resolution. 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), pages 456–465, 2021. 3

    2022

  22. [30]

    Mantiuk, Kil Joong Kim, Allan G

    Rafał K. Mantiuk, Kil Joong Kim, Allan G. Rempel, and Wolfgang Heidrich. Hdr-vdp-2: a calibrated visual metric for visibility and quality predictions in all luminance conditions. SIGGRAPH, 2011. 3

    2011

  23. [31]

    Mantiuk, Alexandre Chapiro, Gizem Rufo, Trisha Lian, Rafał K

    Rafał K. Mantiuk, Alexandre Chapiro, Gizem Rufo, Trisha Lian, Rafał K. Mantiuk, Gyorgy Denes, Alexandre Chapiro, and Anton Kaplanyan. Fovvideovdp: A visible difference predictor for wide field-of-view video. SIGGRAPH, 40:1 – 19, 2021. 4

    2021

  24. [32]

    Mantiuk, Maliha Ashraf, and Alexandre Chapiro

    Rafał K. Mantiuk, Maliha Ashraf, and Alexandre Chapiro. stelacsf: a unified model of contrast sensitivity as the func- tion of spatio-temporal frequency, eccentricity, luminance and area. ACM Trans. Graph., 41(4), 2022. 1

    2022

  25. [33]

    Mantiuk, Param Hanji, Maliha Ashraf, Yuta Asano, and Alexandre Chapiro

    Rafał K. Mantiuk, Param Hanji, Maliha Ashraf, Yuta Asano, and Alexandre Chapiro. Colorvideovdp: A visual differ- ence predictor for image, video and display distortions. SIG- GRAPH, 2024. 4, 6

    2024

  26. [34]

    Srinivasan, Matthew Tancik, Jonathan T

    Ben Mildenhall, Pratul P. Srinivasan, Matthew Tancik, Jonathan T. Barron, Ravi Ramamoorthi, and Ren Ng. Nerf: representing scenes as neural radiance fields for view synthe- sis. Commun. ACM, 65(1):99–106, 2021. 6

    2021

  27. [35]

    Understanding ssim

    Jim Nilsson and Tomas Akenine-M ¨oller. Understanding ssim. arXiv preprint arXiv:2006.13846, 2020. 7, 8

    2006 arXiv

  28. [36]

    Deep learning super sampling (dlss)

    NVIDIA Corporation. Deep learning super sampling (dlss). https : / / www . nvidia . com / en - us / geforce / technologies/dlss/, 2024. Accessed: 2024-11-15. 1, 3

    2024

  29. [37]

    Nvidia tensor core architec- ture

    NVIDIA Corporation. Nvidia tensor core architec- ture. https://www.nvidia.com/en- us/data- center/tensorcore/, 2024. Accessed: 2024-11-15. 1

    2024

  30. [38]

    Limitations of the ssim quality metric in the context of diagnostic imaging

    Jean-Franc ¸ois Pambrun and Rita Noumeir. Limitations of the ssim quality metric in the context of diagnostic imaging. In 2015 IEEE International Conference on Image Processing (ICIP), pages 2960–2963, 2015. 7, 8

    2015

  31. [39]

    Towards foveated rendering for gaze-tracked virtual reality

    Anjul Patney, Marco Salvi, Joohwan Kim, Anton Kaplanyan, Chris Wyman, Nir Benty, David Luebke, and Aaron Lefohn. Towards foveated rendering for gaze-tracked virtual reality. ACM Trans. Graph., 35(6), 2016. 1, 2

    2016

  32. [40]

    Contrast in complex images

    E Peli. Contrast in complex images. J Opt Soc Am A, 7(10): 2032–2040, 1990. 5

    1990

  33. [41]

    High-resolution image syn- thesis with latent diffusion models, 2021

    Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Bj¨orn Ommer. High-resolution image syn- thesis with latent diffusion models, 2021. 3

    2021

  34. [42]

    Laion-5b: an open large-scale dataset for training next generation image-text models

    Christoph Schuhmann, Romain Beaumont, Richard Vencu, Cade Gordon, Ross Wightman, Mehdi Cherti, Theo Coombes, Aarush Katta, Clayton Mullis, Mitchell Worts- man, Patrick Schramowski, Srivatsa Kundurthy, Katherine Crowson, Ludwig Schmidt, Robert Kaczmarczyk, and Jenia Jitsev. Lai...

    2024

  35. [43]

    Robert Shapley, Ehud Kaplan, and Keith P. Purpura. Con- trast sensitivity and light adaptation in photoreceptors or in the retinal network. Contrast Sensitvity (Proceedings of the Retina Research Foundation Symposia), 1993. 3

    1993

  36. [44]

    Adapool: Expo- nential adaptive pooling for information-retaining downsam- pling, 2022

    Alexandros Stergiou and Ronald Poppe. Adapool: Expo- nential adaptive pooling for information-retaining downsam- pling, 2022. 1

    2022

  37. [45]

    Percep- tually adaptive real-time tone mapping

    Taimoor Tariq, Nathan Matsuda, Eric Penner, Jerry Jia, Dou- glas Lanman, Ajit Ninan, and Alexandre Chapiro. Percep- tually adaptive real-time tone mapping. SIGGRAPH Asia,

  38. [46]

    Tursun, Elena Arabadzhiyska-Koleva, Marek Wernikowski, Radosław Mantiuk, Hans-Peter Seidel, Karol Myszkowski, and Piotr Didyk

    O. Tursun, Elena Arabadzhiyska-Koleva, Marek Wernikowski, Radosław Mantiuk, Hans-Peter Seidel, Karol Myszkowski, and Piotr Didyk. Luminance-contrast-aware foveated rendering. SIGGRAPH, 2019. 4, 5

    2019

  39. [47]

    Unsuper- vised degradation representation learning for blind super- resolution

    Longguang Wang, Yingqian Wang, Xiaoyu Dong, Qingyu Xu, Jungang Yang, Wei An, and Yulan Guo. Unsuper- vised degradation representation learning for blind super- resolution. 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 10576–10585, 2021. 3

    2021

  40. [48]

    Adaptive patch exiting for scalable single image super-resolution

    Shizun Wang, Jiaming Liu, Kaixin Chen, Xiaoqi Li, Ming Lu, and Yandong Guo. Adaptive patch exiting for scalable single image super-resolution. In European Conference on Computer Vision, pages 292–307. Springer, 2022. 2

    2022

  41. [49]

    Esrgan: Enhanced super-resolution generative adversarial networks

    Xintao Wang, Ke Yu, Shixiang Wu, Jinjin Gu, Yihao Liu, Chao Dong, Chen Change Loy, Yu Qiao, and Xiaoou Tang. Esrgan: Enhanced super-resolution generative adversarial networks. In ECCV Workshops, 2018. 3

    2018

  42. [50]

    A fully progressive approach to single-image super-resolution

    Yifan Wang, Federico Perazzi, Brian McWilliams, Alexan- der Sorkine-Hornung, Olga Sorkine-Hornung, and Christo- pher Schroers. A fully progressive approach to single-image super-resolution. 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), ...

    2018

  43. [51]

    A formula for human retinal ganglion cell receptive field density as a function of visual field location

    Andrew B Watson. A formula for human retinal ganglion cell receptive field density as a function of visual field location. J Vis, 14(7), 2014. 2

    2014

  44. [52]

    Dif- fir: Efficient diffusion model for image restoration

    Bin Xia, Yulun Zhang, Shiyin Wang, Yitong Wang, Xing Wu, Yapeng Tian, Wenming Yang, and Luc Van Gool. Dif- fir: Efficient diffusion model for image restoration. 2023 IEEE/CVF International Conference on Computer Vision (ICCV), pages 13049–13059, 2023. 3

    2023

  45. [53]

    Learning frequency-aware dy- namic network for efficient super-resolution

    Wenbin Xie, Dehua Song, Chang Xu, Chunjing Xu, Hui Zhang, and Yunhe Wang. Learning frequency-aware dy- namic network for efficient super-resolution. In Proceedings of the IEEE/CVF International Conference on Computer Vi- sion, pages 4308–4317, 2021. 2

    2021

  46. [54]

    Image super-resolution as sparse representation of raw image patches

    Jianchao Yang, John Wright, Thomas Huang, and Yi Ma. Image super-resolution as sparse representation of raw image patches. In 2008 IEEE Conference on Computer Vision and Pattern Recognition, pages 1–8, 2008. 3

    2008

  47. [55]

    Daly, and Shawmin Lei

    Wenjun Kevin Zeng, Scott J. Daly, and Shawmin Lei. Point- wise extended visual masking for jpeg-2000 image compres- sion. International Conference on Image Processing (ICIP), 1:657–660 vol.1, 2000. 5 10

    2000

  48. [56]

    On single image scale-up using sparse-representations

    Roman Zeyde, Michael Elad, and Matan Protter. On single image scale-up using sparse-representations. In Curves and Surfaces, pages 711–730, Berlin, Heidelberg, 2012. Springer Berlin Heidelberg. 5

    2012

  49. [57]

    Efros, Eli Shecht- man, and Oliver Wang

    Richard Zhang, Phillip Isola, Alexei A. Efros, Eli Shecht- man, and Oliver Wang. The unreasonable effectiveness of deep features as a perceptual metric. 2018 IEEE/CVF Con- ference on Computer Vision and Pattern Recognition, pages 586–595, 2018. 3

    2018

  50. [58]

    The unreasonable effectiveness of deep features as a perceptual metric

    Richard Zhang, Phillip Isola, Alexei A Efros, Eli Shecht- man, and Oliver Wang. The unreasonable effectiveness of deep features as a perceptual metric. In Proceedings of the IEEE conference on computer vision and pattern recogni- tion, pages 586–595, 2018. 7

    2018

  51. [59]

    Theoretical understanding of the early visual processes by data compression and data selection

    Li Zhaoping. Theoretical understanding of the early visual processes by data compression and data selection. Network: computation in neural systems, 17(4):301–334, 2006. 1, 3 11 Human Vision Constrained Super-Resolution Supplementary Material Abstract This supplementary file p...

    2006

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.