Pith. sign in

REVIEW 3 major objections 5 minor 43 references

RankSRGAN: Generative Adversarial Networks with Ranker for Image Super-Resolution

T0 review · 3 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read A learned Ranker lets super-resolution GANs optimize perceptual scores directly, and the paper reports best NIQE/PI on three benchmarks.

desk verdict A genuine, well-ablated approach to optimizing non-differentiable perceptual metrics in SR; the main gap is how the frozen Ranker behaves on the generator's shifting output distribution. read the letter →

arxiv 1908.06382 v2 pith:FBWKQTQ7 submitted 2019-08-18 cs.CV

Wenlong Zhang , Yihao Liu , Chao Dong , Yu Qiao This is my paper · ORCID
classification cs.CV
keywords RankSRGANsingleimagesuper-resolutionperceptualmetricslearningtoSiamesenetworkgenerativeadversarialNIQEno-referencequalityassessment
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

Perceptual super-resolution is judged by no-reference metrics such as NIQE, Ma, and PI, but these metrics are not differentiable, so they cannot directly train a generator. RankSRGAN adds a Ranker, a Siamese CNN trained with margin-ranking loss on pairs of super-resolved images labeled by the chosen metric, to supply a differentiable rank-content loss. The generator is then optimized toward higher perceptual scores while keeping the standard SRGAN objective. The authors show this lets a standard SRGAN beat SRGAN and ESRGAN on NIQE and PI on Set14, BSD100, and PIRM-Test, and that the framework generalizes across metrics and rank datasets.

What carries the argument

The Ranker is a Siamese CNN that takes two images and outputs scalar ranking scores, trained with margin-ranking loss to match the pair-wise order of a chosen perceptual metric. It converts an indifferentiable metric into a differentiable objective; the concrete hook is the rank-content loss $L_R = \mathrm{sigmoid}(R(G(x_i)))$, which lets gradients flow from the predicted rank back into the generator. The metric-rank labeling scheme, rather than model classification, is what lets the framework exceed the upper bound of any single SR algorithm.

What would settle it

Take a trained RankSRGAN and compare its Ranker scores against the true perceptual metric on the generator's own outputs: if the Spearman correlation drops well below the reported validation value of 0.88, or if a model trained with the rank-content loss improves its Ranker score without improving the actual metric, the Ranker has stopped tracking the target.

Watch

Extended reading notes

Core claim

The central claim is that any non-differentiable perceptual metric can be optimized through a learned ranking surrogate. The Ranker is a Siamese VGG-style network trained to reproduce the pair-wise ordering induced by a metric on outputs of several SR algorithms; training uses margin-ranking loss rather than regression, so the surrogate only needs ordinal agreement. The rank-content loss $L_R = \mathrm{sigmoid}(R(G(x_i)))$ is added to the standard SRGAN objective, and lowering it raises the predicted perceptual rank. Because the rank dataset contains images where SRGAN and ESRGAN each win in mixed order, the generator is guided to adopt the better parts of either algorithm. The reported experiments give the best NIQE and PI among the compared methods on Set14, BSD100, and PIRM-Test, preserve PSNR on PIRM-Test, and user studies prefer RankSRGAN over both baselines.

Load-bearing premise

The Ranker is trained only on super-resolved outputs from SRResNet, SRGAN, and ESRGAN under a fixed perceptual metric, and the method assumes its ranking signal stays valid for the generator's changing output distribution during adversarial training.

Editorial extensions

If this is right

  • Swapping the labels in the rank dataset swaps the target metric: RankSRGAN trained with NIQE, Ma, or PI labels improves that metric, so the same framework can chase any pair-wise rankable quality score.
  • Because the Ranker learns order rather than absolute values, it separates SR algorithms whose metric values are close (for example SRGAN vs ESRGAN on NIQE), giving the generator a usable gradient where regression would collapse.
  • The metric-rank upper bound can exceed every SR algorithm in the rank dataset, so adding better or complementary guiding algorithms, even ground-truth HR, should continue to improve perceptual quality.
  • The rank-content loss can be weighted against MSE to navigate the perception-distortion tradeoff, letting practitioners choose a preferred balance between visual quality and fidelity.
  • The generator's parameter count and inference cost stay the same as SRGAN; only training time grows by about 1.18 times.

Reading between the lines

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

  • An unstated consequence is that any non-differentiable evaluation score expressible as pair-wise preferences, not just NR-IQA metrics, could be optimized with the same Ranker mechanism, including full-reference quality measures or task-specific scores.
  • A risk the paper does not address is reward hacking: the generator may exploit the Ranker's learned scoring function rather than the true metric if training pushes outputs outside the rank dataset's distribution, and the reported validation SROCC of 0.88 does not guarantee the same ordinal accuracy on the generator's own outputs.
  • A testable extension is to build rank datasets from more diverse or deliberately adversarial SR models, or to update the Ranker periodically during GAN training, to see whether the ranking signal stays aligned with the target metric under distribution shift.
  • The metric-rank versus model-classification comparison suggests an ensemble view: a Ranker over many specialist generators could yield a single model that inherits each specialist's best behavior per image.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. RankSRGAN proposes a super-resolution framework that augments SRGAN with a learned Ranker, a Siamese CNN trained with margin-ranking loss to mimic the ordering behavior of non-differentiable perceptual metrics such as NIQE, PI, and Ma. The Ranker is trained on rank datasets constructed from the outputs of SRResNet, SRGAN, and ESRGAN on DIV2K/Flickr2K. Once trained, the Ranker's scalar output is used to define a rank-content loss, and the generator is optimized by a weighted sum of perceptual, adversarial, and rank-content losses. Experiments on Set14, BSD100, and PIRM-Test report lower NIQE/PI than SRGAN and ESRGAN, with ablations on the rank dataset composition, the choice of perceptual metric, ranking versus regression for the Ranker, and the loss combination.

Significance. If the mechanism is sound, the idea of using a learned ranker as a differentiable surrogate for non-differentiable perceptual quality metrics is a useful addition to perceptual SR, and the paper demonstrates that this can combine the strengths of different SR algorithms. The ablations (rank vs. regression, different metrics, different rank datasets) are thoughtful and provide some evidence that the Ranker is not merely a black-box regularizer. The paper also reports SROCC of the Ranker and includes convergence curves, which is helpful. However, the central claim that the rank-content loss 'optimizes the generator in the direction of perceptual metrics' rests on an unvalidated assumption: that the Ranker's absolute score is a valid loss signal for images that fall outside its training distribution. The empirical gains are small and reported without error bars or significance tests, so the strength of the claim currently exceeds the evidence.

major comments (3)
  1. [Sec. 3.3, Eq. (9)] The rank-content loss in Eq. (9) is the sigmoid of the Ranker's raw output, but the Ranker is trained with a margin-ranking loss (Eq. (4)) that determines scores only up to an arbitrary monotonic transformation. Without calibration or a mechanism anchoring the absolute scale of R to the target perceptual metric, minimizing sigmoid(R(G(x))) is not theoretically guaranteed to minimize the metric. The paper should provide either a calibration step or empirical evidence that lower R(G(x)) consistently corresponds to better metric values on a validation set of images produced during training.
  2. [Sec. 4.1-4.2] The Ranker is trained only on outputs of SRResNet, SRGAN, and ESRGAN, and its accuracy is evaluated only via SROCC on a held-out split of that same rank dataset. During adversarial training, the generator's output distribution shifts (as seen in Figure 6), yet the paper does not measure how the Ranker's ranking accuracy behaves on the generator's intermediate or final outputs. Without this distribution-shift analysis, the key claim that the rank-content loss 'optimizes the generator in the direction of perceptual metrics' is not fully supported; it is possible that the Ranker acts as a regularizer that only biases outputs toward the training SR methods. Please evaluate the Ranker on generated images from the trained RankSRGAN and report the correlation between R(G(x)) and the metric.
  3. [Tables 2-7] All quantitative results are reported as single-run means without error bars or significance tests. Given that the improvements over ESRGAN are often within a few hundredths of a NIQE/PI point (e.g., Table 2, Set14: NIQE 3.28 vs 3.28, PI 2.61 vs 2.61), it is unclear whether the claimed gains are statistically meaningful. I request variance across at least three independent training runs or a statistical significance test on the test datasets.
minor comments (5)
  1. [Sec. 4.1, Eq. (11)] The SROCC definition is written only for the case of no tied ranks; please clarify how ties are handled or state that the data contain no ties.
  2. [Sec. 3.4, Eq. (10)] The notation 'PMSR2-L + PMSR2-H' and 'PMSR1-L' is confusing; the subscripts should be defined explicitly.
  3. [Figure 3] The caption does not match the bars; please indicate which bar corresponds to 'Metric Rank', 'Model Classification', 'ESRGAN', and 'SRGAN'.
  4. [Sec. 4.4, Table 7] The values of alpha_1 and alpha_2 are given as '{1,5}' but it is unclear which alpha corresponds to which row; please label the rows explicitly.
  5. [Throughout] There are minor typos: 'beheviour' in Section 3.1, 'traning' in Section 4.2, and 'indifferentiable' should likely be 'non-differentiable' in the introduction.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the Ranker is a learned surrogate, while final evaluation uses the original metrics on held-out sets.

full rationale

The paper's central claim is that RankSRGAN optimizes a generator in the direction of otherwise non-differentiable perceptual metrics by training a Ranker to imitate metric rankings and then using the Ranker output as a loss. This is a surrogate-optimization procedure, not a circular derivation. The Ranker is trained on outputs of SRResNet, SRGAN, and ESRGAN with labels from NIQE, Ma, or PI, and then the rank-content loss (Eq. 9) minimizes sigmoid of the Ranker score. If the Ranker perfectly mimicked the metric, minimizing it would indeed improve the metric, but the paper does not define the target metric as the Ranker. The final evaluations in Table 2 use the original NIQE and PI values computed directly on Set14, BSD100, and PIRM-Test, which are external to the Ranker training data (DIV2K/Flickr2K). Thus the reported improvements are not forced by construction; they depend on whether the Ranker generalizes. The Ranker's SROCC of 0.88 is reported on a held-out split of the rank dataset, which is evidence of ordering ability but does not by itself guarantee distribution-shift robustness. That is a correctness or generalization concern, not a circularity. The paper also cites ESRGAN, which shares overlapping authors, but ESRGAN is used as an external baseline and data source, not as an unverified premise that carries the argument. No equation in the paper reduces to its own inputs, and no fitted parameter is relabeled as a prediction. The upper-bound analysis in Section 3.4 is descriptive and is separately validated by experiments comparing RankSRGAN-MR and RankSRGAN-MC. Overall, the derivation chain is self-contained with respect to the tested metrics; the main limitations are empirical robustness issues rather than circular reasoning.

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

The paper's central mechanism adds a learned ranking network on top of a standard SRGAN. It introduces no physical entities and no first-principles constants; the main fitted quantities are training hyperparameters. The key domain assumptions are that the chosen perceptual metrics reflect human quality judgments and that a Siamese network trained on outputs of three SR methods generalizes as a ranker.

free parameters (3)
  • margin epsilon in margin-ranking loss = 0.5
    Set by hand in Eq. 4 to control the distance between ranking scores; affects Ranker training.
  • adversarial loss weight in Eq. 12 = 0.005
    Determined empirically to balance GAN loss against rank-content and perceptual losses.
  • rank-content loss weight in Eq. 12 = 0.03
    Determined empirically; Table 7 shows that changing MSE loss weights shifts NIQE/PSNR trade-off, indicating sensitivity.
assumptions (4)
  • domain assumption NIQE, PI and Ma are valid proxies for human perceptual quality.
    Paper relies on correlations reported in PIRM2018-SR challenge [2] (NIQE 0.76, PI 0.83) to justify optimizing these metrics.
  • domain assumption A Siamese CNN trained with margin-ranking loss can learn the ordering behavior of a non-differentiable perceptual metric.
    Central to the method; validated only by the reported SROCC 0.88 on a validation split, not by an external benchmark.
  • domain assumption The rank dataset built from SRResNet, SRGAN and ESRGAN outputs covers the relevant distribution of perceptual quality for training.
    The generator is trained against this Ranker, so the Ranker's reliability on out-of-distribution images is assumed.
  • domain assumption The SRGAN base model and adversarial training setup are reliable.
    The framework builds directly on SRGAN [22]; no new theoretical justification for the GAN objective is given.

how reviews work

0 comments
Cite this review

Pith. "Pith review of RankSRGAN: Generative Adversarial Networks with Ranker for Image Super-Resolution." pith.science (2026). https://pith.science/paper/FBWKQTQ7

@misc{pith2026190806382,
  author       = {Pith},
  title        = {Pith review of: RankSRGAN: Generative Adversarial Networks with Ranker for Image Super-Resolution},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FBWKQTQ7}},
  note         = {Machine review of arXiv:1908.06382}
}
read the original abstract

Generative Adversarial Networks (GAN) have demonstrated the potential to recover realistic details for single image super-resolution (SISR). To further improve the visual quality of super-resolved results, PIRM2018-SR Challenge employed perceptual metrics to assess the perceptual quality, such as PI, NIQE, and Ma. However, existing methods cannot directly optimize these indifferentiable perceptual metrics, which are shown to be highly correlated with human ratings. To address the problem, we propose Super-Resolution Generative Adversarial Networks with Ranker (RankSRGAN) to optimize generator in the direction of perceptual metrics. Specifically, we first train a Ranker which can learn the behavior of perceptual metrics and then introduce a novel rank-content loss to optimize the perceptual quality. The most appealing part is that the proposed method can combine the strengths of different SR methods to generate better results. Extensive experiments show that RankSRGAN achieves visually pleasing results and reaches state-of-the-art performance in perceptual metrics. Project page: https://wenlongzhang0724.github.io/Projects/RankSRGAN

Figures

Figures reproduced from arXiv: 1908.06382 by the authors.

Figure 1
Figure 1. The comparison of RankSRGAN and the state-of-the [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Overview of the proposed method. Stage 1: Generate pair-wise rank images by different SR models in the orientation of perceptual metrics. Stage 2: Train Siamese-like Ranker network. Stage 3: Introduce rank-content loss derived from well-trained Ranker to guide GAN training. RankSRGAN consists of a generator(G), discriminator(D), a fixed Feature extractor(F) and Ranker(R). 3. Method 3.1. Overview of RankSRGAN The pro… view at source ↗
Figure 3
Figure 3. The upper bound (average NIQE value) of SRGAN, [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figures from the paper (10 more)
Figure 5
Figure 5. Figure 5: Visual results comparison of our model with other works on [PITH_FULL_IMAGE:figures/full_fig_p006_5.png]
Figure 6
Figure 6. Figure 6: Convergence curves of RankSRGAN in PSNR/ NIQE. [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]
Figure 7
Figure 7. Figure 7: The results of user studies, comparing our method with [PITH_FULL_IMAGE:figures/full_fig_p008_7.png]
Figure 8
Figure 8. Figure 8: The ranking results of user studies: SRResNet [ [PITH_FULL_IMAGE:figures/full_fig_p008_8.png]
Figure 1
Figure 1. Figure 1: The convergence curves of RankSRGAN with Ranker1 and Ranker2 in NIQE and PSNR. [PITH_FULL_IMAGE:figures/full_fig_p011_1.png]
Figure 2
Figure 2. Figure 2: The convergence curves of RankSRGAN with Ranker-VGG-8 and Ranker-VGG-12 in NIQE and PSNR. [PITH_FULL_IMAGE:figures/full_fig_p012_2.png]
Figure 3
Figure 3. Figure 3: The histograms of (a) NIQE label value, (b) the regression score of “regression” Ranker and (c) the ranking score of Ranker [PITH_FULL_IMAGE:figures/full_fig_p013_3.png]
Figure 4
Figure 4. Figure 4: The histograms of (a) Ma label value, (b) the regression score of “regression” Ranker and (c) the ranking score of Ranker (ours). [PITH_FULL_IMAGE:figures/full_fig_p013_4.png]
Figure 5
Figure 5. Figure 5: The convergence curves of RankSRGANN , RankSRGANM and RankSRGANPI . 2.2. Convergence curves for RankSRGAN-HR As shown in [PITH_FULL_IMAGE:figures/full_fig_p014_5.png]
Figure 6
Figure 6. Figure 6: The convergence curves of RankSRGAN-HR in PSNR and NIQE. [PITH_FULL_IMAGE:figures/full_fig_p014_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

43 extracted references · 36 canonical work pages

  1. [1]

    Ntire 2017 challenge on single image super-resolution: Dataset and study

    Eirikur Agustsson and Radu Timofte. Ntire 2017 challenge on single image super-resolution: Dataset and study. In The IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Workshops , volume 3, page 2, 2017

    work page 2017

  2. [2]

    2018 pirm challenge on perceptual image super-resolution

    Yochai Blau, Roey Mechrez, Radu Timofte, Tomer Michaeli, and Lihi Zelnik-Manor. 2018 pirm challenge on perceptual image super-resolution. arXiv preprint arXiv:1809.07517 , 2018

    arXiv 2018

  3. [3]

    The perception-distortion tradeoff

    Yochai Blau and Tomer Michaeli. The perception-distortion tradeoff. In Proc. 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, Salt Lake City, Utah, USA , pages 6228--6237, 2018

    work page 2018

  4. [4]

    Deep neural networks for no-reference and full-reference image quality assessment

    Sebastian Bosse, Dominique Maniry, Klaus-Robert M \"u ller, Thomas Wiegand, and Wojciech Samek. Deep neural networks for no-reference and full-reference image quality assessment. IEEE Transactions on Image Processing , 27(1):206--219, 2018

    work page 2018

  5. [5]

    a ckinger, and Roopak Shah. Signature verification using a

    Jane Bromley, Isabelle Guyon, Yann LeCun, Eduard S \"a ckinger, and Roopak Shah. Signature verification using a" siamese" time delay neural network. In Advances in neural information processing systems , pages 737--744, 1994

    work page 1994

  6. [6]

    Super-resolution with deep convolutional sufficient statistics

    Joan Bruna, Pablo Sprechmann, and Yann LeCun. Super-resolution with deep convolutional sufficient statistics. arXiv preprint arXiv:1511.05666 , 2015

    arXiv 2015

  7. [7]

    Learning to rank using gradient descent

    Chris Burges, Tal Shaked, Erin Renshaw, Ari Lazier, Matt Deeds, Nicole Hamilton, and Greg Hullender. Learning to rank using gradient descent. In Proceedings of the 22nd international conference on Machine learning , pages 89--96. ACM, 2005

    work page 2005

  8. [8]

    Deep Learning-based Image Super-Resolution Considering Quantitative and Perceptual Quality

    Jun-Ho Choi, Jun-Hyuk Kim, Manri Cheon, and Jong-Seok Lee. Deep learning-based image super-resolution considering quantitative and perceptual quality. arXiv preprint arXiv:1809.04789 , 2018

    work page Pith review arXiv 2018

Show all 43 references
  1. [9]

    Learning a similarity metric discriminatively, with application to face verification

    Sumit Chopra, Raia Hadsell, and Yann LeCun. Learning a similarity metric discriminatively, with application to face verification. In Computer Vision and Pattern Recognition, 2005. CVPR 2005. IEEE Computer Society Conference on , volume 1, pages 539--546. IEEE, 2005

    2005

  2. [10]

    Learning a deep convolutional network for image super-resolution

    Chao Dong, Chen Change Loy, Kaiming He, and Xiaoou Tang. Learning a deep convolutional network for image super-resolution. In European conference on computer vision , pages 184--199. Springer, 2014

    2014

  3. [11]

    Accelerating the super-resolution convolutional neural network

    Chao Dong, Chen Change Loy, and Xiaoou Tang. Accelerating the super-resolution convolutional neural network. In European Conference on Computer Vision , pages 391--407. Springer, 2016

    2016

  4. [12]

    Generating images with perceptual similarity metrics based on deep networks

    Alexey Dosovitskiy and Thomas Brox. Generating images with perceptual similarity metrics based on deep networks. In Advances in Neural Information Processing Systems , pages 658--666, 2016

    2016

  5. [13]

    Suppressing model overfitting for image super-resolution networks

    Ruicheng Feng, Jinjin Gu, Yu Qiao, and Chao Dong. Suppressing model overfitting for image super-resolution networks. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops , pages 0--0, 2019

    2019

  6. [14]

    Blind super-resolution with iterative kernel correction

    Jinjin Gu, Hannan Lu, Wangmeng Zuo, and Chao Dong. Blind super-resolution with iterative kernel correction. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , pages 1604--1613, 2019

    2019

  7. [15]

    Deep backprojection networks for super-resolution

    Muhammad Haris, Greg Shakhnarovich, and Norimichi Ukita. Deep backprojection networks for super-resolution. In Conference on Computer Vision and Pattern Recognition , 2018

    2018

  8. [16]

    Modulating image restoration with continual levels via adaptive feature modification layers

    Jingwen He, Chao Dong, and Yu Qiao. Modulating image restoration with continual levels via adaptive feature modification layers. In The IEEE Conference on Computer Vision and Pattern Recognition (CVPR) , June 2019

    2019

  9. [17]

    Delving deep into rectifiers: Surpassing human-level performance on imagenet classification

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Delving deep into rectifiers: Surpassing human-level performance on imagenet classification. In Proceedings of the IEEE international conference on computer vision , pages 1026--1034, 2015

    2015

  10. [18]

    Perceptual losses for real-time style transfer and super-resolution

    Justin Johnson, Alexandre Alahi, and Li Fei-Fei. Perceptual losses for real-time style transfer and super-resolution. In European Conference on Computer Vision , pages 694--711. Springer, 2016

    2016

  11. [19]

    Convolutional neural networks for no-reference image quality assessment

    Le Kang, Peng Ye, Yi Li, and David Doermann. Convolutional neural networks for no-reference image quality assessment. In Proceedings of the IEEE conference on computer vision and pattern recognition , pages 1733--1740, 2014

    2014

  12. [20]

    Accurate image super-resolution using very deep convolutional networks

    Jiwon Kim, Jung Kwon Lee, and Kyoung Mu Lee. Accurate image super-resolution using very deep convolutional networks. In Proceedings of the IEEE conference on computer vision and pattern recognition , pages 1646--1654, 2016

    2016

  13. [21]

    Adam: A method for stochastic optimization

    Diederik P Kingma and Jimmy Ba. Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980 , 2014

    2014 arXiv

  14. [22]

    Photo-realistic single image super-resolution using a generative adversarial network

    Christian Ledig, Lucas Theis, Ferenc Husz \'a r, Jose Caballero, Andrew Cunningham, Alejandro Acosta, Andrew P Aitken, Alykhan Tejani, Johannes Totz, Zehan Wang, et al. Photo-realistic single image super-resolution using a generative adversarial network. In CVPR , volume 2, pa...

    2017

  15. [23]

    Blind image quality assessment using a general regression neural network

    Chaofeng Li, Alan Conrad Bovik, and Xiaojun Wu. Blind image quality assessment using a general regression neural network. IEEE Transactions on Neural Networks , 22(5):793--799, 2011

    2011

  16. [24]

    Enhanced deep residual networks for single image super-resolution

    Bee Lim, Sanghyun Son, Heewon Kim, Seungjun Nah, and Kyoung Mu Lee. Enhanced deep residual networks for single image super-resolution. In The IEEE conference on computer vision and pattern recognition (CVPR) workshops , volume 1, page 4, 2017

    2017

  17. [25]

    Rankiqa: Learning from rankings for no-reference image quality assessment

    Xialei Liu, Joost van de Weijer, and Andrew D Bagdanov. Rankiqa: Learning from rankings for no-reference image quality assessment. Computer Vision and Pattern Recognition, https://arxiv. org/abs/1707.08347 v1 , 2017

    2017 arXiv

  18. [26]

    Learning a no-reference quality metric for single-image super-resolution

    Chao Ma, Chih-Yuan Yang, Xiaokang Yang, and Ming-Hsuan Yang. Learning a no-reference quality metric for single-image super-resolution. Computer Vision and Image Understanding , 158:1--16, 2017

    2017

  19. [27]

    A database of human segmented natural images and its application to evaluating segmentation algorithms and measuring ecological statistics

    David Martin, Charless Fowlkes, Doron Tal, and Jitendra Malik. A database of human segmented natural images and its application to evaluating segmentation algorithms and measuring ecological statistics. In Computer Vision, 2001. ICCV 2001. Proceedings. Eighth IEEE Internationa...

    2001

  20. [28]

    Conditional generative adversarial nets

    Mehdi Mirza and Simon Osindero. Conditional generative adversarial nets. arXiv preprint arXiv:1411.1784 , 2014

    2014 arXiv

  21. [29]

    completely blind

    Anish Mittal, Rajiv Soundararajan, and Alan C Bovik. Making a" completely blind" image quality analyzer. IEEE Signal Process. Lett. , 20(3):209--212, 2013

    2013

  22. [30]

    Relative attributes

    Devi Parikh and Kristen Grauman. Relative attributes. In Computer Vision (ICCV), 2011 IEEE International Conference on , pages 503--510. IEEE, 2011

    2011

  23. [31]

    Enhancenet: Single image super-resolution through automated texture synthesis

    Mehdi SM Sajjadi, Bernhard Sch \"o lkopf, and Michael Hirsch. Enhancenet: Single image super-resolution through automated texture synthesis. In Computer Vision (ICCV), 2017 IEEE International Conference on , pages 4501--4510. IEEE, 2017

    2017

  24. [32]

    Ranking cgans: Subjective control over semantic image attributes

    Yassir Saquil, Kwang In Kim, and Peter Hall. Ranking cgans: Subjective control over semantic image attributes. In Proc. of British Machine Vision Conference (BMVC) , 2018

    2018

  25. [33]

    Very deep convolutional networks for large-scale image recognition

    Karen Simonyan and Andrew Zisserman. Very deep convolutional networks for large-scale image recognition. arXiv preprint arXiv:1409.1556 , 2014

    2014 arXiv

  26. [34]

    Recovering realistic texture in image super-resolution by deep spatial feature transform

    Xintao Wang, Ke Yu, Chao Dong, and Chen Change Loy. Recovering realistic texture in image super-resolution by deep spatial feature transform. arXiv preprint arXiv:1804.02815 , 2018

    2018 arXiv

  27. [35]

    Esrgan: Enhanced super-resolution generative adversarial networks

    Xintao Wang, Ke Yu, Shixiang Wu, Jinjin Gu, Yihao Liu, Chao Dong, Yu Qiao, and Chen Change Loy. Esrgan: Enhanced super-resolution generative adversarial networks. In The European Conference on Computer Vision (ECCV) Workshops , September 2018

    2018

  28. [36]

    Deep relative attributes

    Xiaoshan Yang, Tianzhu Zhang, Changsheng Xu, Shuicheng Yan, M Shamim Hossain, and Ahmed Ghoneim. Deep relative attributes. IEEE Transactions on Multimedia , 18(9):1832--1842, 2016

    2016

  29. [37]

    No-reference image quality assessment based on visual codebook

    Peng Ye and David S Doermann. No-reference image quality assessment based on visual codebook. In ICIP , pages 3089--3092. Citeseer, 2011

    2011

  30. [38]

    Learning to compare image patches via convolutional neural networks

    Sergey Zagoruyko and Nikos Komodakis. Learning to compare image patches via convolutional neural networks. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition , pages 4353--4361, 2015

    2015

  31. [39]

    On single image scale-up using sparse-representations

    Roman Zeyde, Michael Elad, and Matan Protter. On single image scale-up using sparse-representations. In International conference on curves and surfaces , pages 711--730. Springer, 2010

    2010

  32. [40]

    Single image super-resolution with non-local means and steering kernel regression

    Kaibing Zhang, Xinbo Gao, Dacheng Tao, Xuelong Li, et al. Single image super-resolution with non-local means and steering kernel regression. Image , 11:12, 2012

    2012

  33. [41]

    An edge-guided image interpolation algorithm via directional filtering and data fusion

    Lei Zhang and Xiaolin Wu. An edge-guided image interpolation algorithm via directional filtering and data fusion. IEEE transactions on Image Processing , 15(8):2226--2238, 2006

    2006

  34. [42]

    Residual dense network for image super-resolution

    Yulun Zhang, Yapeng Tian, Yu Kong, Bineng Zhong, and Yun Fu. Residual dense network for image super-resolution. In The IEEE Conference on Computer Vision and Pattern Recognition (CVPR) , 2018

    2018

  35. [43]

    write newline

    " write newline "" before.all 'output.state := FUNCTION fin.entry add.period write newline FUNCTION new.block output.state before.all = 'skip after.block 'output.state := if FUNCTION new.sentence output.state after.block = 'skip output.state before.all = 'skip after.sentence '...

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.