Virtual Thin Slice: 3D Conditional GAN-based Super-resolution for CT Slice Interval

Akira Kudo; Edgar Simo-Serra; Satoshi Iizuka; Yoshiro Kitamura; Yuanzhong Li

REVIEW 3 major objections 4 minor 15 references

Virtual Thin Slice: 3D Conditional GAN-based Super-resolution for CT Slice Interval

T0 review · 3 major / 4 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 conditioned 3D GAN can turn thick-slice CT archives into 1 mm virtual thin slices.

desk verdict A plausible 3D cGAN for CT z-axis super-resolution with a useful conditioning trick, but the clinical value rests on a synthetic degradation model that the paper never validates on real paired data. read the letter →

arxiv 1908.11506 v2 pith:3A4IEL72 submitted 2019-08-30 eess.IV cs.CV

Akira Kudo , Yoshiro Kitamura , Yuanzhong Li , Satoshi Iizuka , Edgar Simo-Serra This is my paper · ORCID

classification eess.IVcs.CV

keywords computedtomographysliceintervalsuper-resolutionconditionalgenerativeadversarialnetwork3Dconvolutionalmodecollapsevolumerenderingvirtualthin

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

Hospitals routinely store CT scans with thick slice intervals to save space, but the resulting volumes are blurry along the body axis and poorly suited to 3D visualization or analysis. This paper claims that a conditional generative adversarial network can convert those archived thick-slice volumes into 1 mm “virtual thin slices” that reproduce bone boundaries and blood vessels sharply enough for expert viewers to prefer them. The proposed Virtual Thin Slice method uses one 3D fully convolutional generator for head, chest, abdomen, and leg CT, and conditions only the discriminator on body part, slice interval, and blur level to counter mode collapse. On 53 test volumes it reports better PSNR and SSIM than bicubic, SRCNN, and a 3D Pix2Pix baseline, and radiology technicians and medical imaging scientists preferred its output roughly 90 percent of the time. The practical stake is that existing thick-slice CT archives could be upgraded without rescanning patients.

What carries the argument

The load-bearing mechanism is the paired design of a 3D fully convolutional generator and a conditioned discriminator. The generator follows a U-Net encoder-decoder with 4×4×4 convolutions, batch normalization, LeakyReLU activation, trilinear upsampling instead of transposed convolutions, and a residual connection that adds the predicted high-frequency component to the input volume. The discriminator receives the thick image, the generated or real thin image, and an 8-channel conditioning vector built from body part (head, chest, abdomen, or leg), slice interval (4 or 8 mm), and the Gaussian sigma level used in degradation, with a self-attention layer inserted in the fourth layer to speed convergence. Training relies on a Degrader that blurs a thin volume with Gaussian smoothing (sigma from 0.0 to 3.2 voxels), downsamples by 1/4 or 1/8, and applies spline interpolation plus random noise; random 160-cube crops with affine augmentation teach the network to handle varying anatomy and fields of view. The conditioning is the invention: it injects into the discriminator the very factors that vary across CT studies, so the adversarial game cannot collapse to a single dominant body part or slice thickness.

What would settle it

Take real CT studies that have both the archived thick-slice reconstruction (3–10 mm slice interval) and the original 1 mm thin-slice reconstruction for the same patient, feed only the thick volume to VTS, and compare the output to the true thin volume with PSNR and SSIM; if the margin over bicubic interpolation disappears on such real pairs, the synthetic Degrader assumption is the point of failure.

Watch

Extended reading notes

Core claim

The paper's central claim is that a conditional GAN, trained entirely on thin-slice CT volumes artificially degraded to mimic thick slices, can generate 1 mm virtual thin slices from 8 mm inputs across four body regions with a single model. The proposed mechanism is to condition the discriminator on an 8-channel vector encoding body part, slice interval, and Gaussian blur level, while leaving the generator unconditional; this is what lets the model keep output variety and avoid mode collapse. The generator is a 3D fully convolutional U-Net that predicts a high-frequency residual added to the input, so it accepts arbitrary fields of view and slice counts. On 53 test volumes, VTS reports PSNR 35.73 and SSIM 0.933, above 3D Pix2Pix (35.14, 0.925), SRCNN (33.73, 0.904), and bicubic (32.34, 0.878), and a Visual Turing Test preferred VTS roughly 90 percent of the time. The paper also verifies the generator on 66 real thick-slice studies with varied spacing and field of view, producing whole-body reconstructions without visible seams.

Load-bearing premise

The training's simulated degradation, combining Gaussian blur, downsampling by a factor of 4 or 8, spline interpolation, and random noise, is assumed to faithfully stand in for how real CT scanners produce thick-slice images; if real thick-slice acquisition physics differ, the learned mapping may not transfer.

Editorial extensions

If this is right

Archived thick-slice CT studies, including whole-body volumes, can be converted to 1 mm-equivalent slices, enabling 3D volume rendering and sagittal or coronal viewing without re-scanning patients.
A single trained network covers head, chest, abdomen, and legs, so hospitals do not need per-anatomy models; the conditioning information is only needed during training and not at test time.
Sharp reconstruction of high-intensity structures such as vertebrae and blood vessels should make tasks like vertebral numbering, bone labeling, and lung segmentation easier on old archives.
Because PSNR and SSIM improve over the baselines and expert raters prefer the output, the method offers a practical way to raise the visualization and analysis value of existing PACS storage.

Reading between the lines

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

If the Degrader assumption holds across different scanner reconstruction kernels, the same architecture could be retrained for other anisotropic volume modalities, such as MRI, by swapping the degradation model and encoding acquisition parameters in the conditioning vector.
The large Visual Turing Test preference alongside modest PSNR gains is consistent with the perception-distortion tradeoff; one concrete prediction is that VTS outputs will score relatively better on perceptual metrics than on pixel-error metrics.
A direct clinical deployment test would be to run VTS on a hospital's archived thick-slice studies and check whether radiologists can perform 3D tasks, such as vertebra labeling or vessel tracking, as reliably on virtual thin slices as on true thin slices; the paper identifies this as future work.

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.

Desk Editor's Note

read the letter

You should know two things: the paper has a genuinely useful architectural idea, and its main evidence rests on a synthetic degradation that may not faithfully reflect real thick-slice CT. The core claim is plausible but not yet proven for clinical data.

What's new: Kudo et al. extend cGAN-based super-resolution to 3D CT slices, with a fully convolutional generator that can handle arbitrary fields of view and a discriminator conditioned on body part, slice interval, and blur sigma. The conditioning is a sensible way to mitigate mode collapse across heterogeneous anatomy, and the ablation shows it matters (PSNR 35.73 vs 35.17). They also use residual high-frequency prediction, which is a nice touch and clearly helps. As far as I can tell, the combination is not in the cited prior work; Chen et al. only did brain MRI.

What's done well: the training pipeline is carefully described—random degradation with Gaussian blur and downsampling, data augmentation with affine transforms, and a clean split of 354/53 volumes. The comparison against bicubic, SRCNN, and Pix2Pix is standard, and they report both PSNR/SSIM and a Visual Turing Test. The VTT, though small, at least uses radiology technicians and medical imaging scientists. The qualitative results on real thick slices are visually appealing, and the fact that the generator runs on arbitrary sizes is practically useful.

Soft spots: the differences in PSNR/SSIM are small—0.59 dB and 0.008 SSIM over Pix2Pix—and no error bars or significance tests are given. That's a minor-to-moderate issue. More important, the Degrader that generates thick slices is a simplified model: Gaussian blur with sigma up to 3.2 voxels, downsample to 1/4 or 1/8, spline interpolation, and random noise. Real CT thick-slice acquisition involves the slice sensitivity profile, reconstruction kernels, and helical interpolation; the paper's model may not capture those. The training data deliberately exclude metal artifacts and noise, so the method's robustness to clinical reality is untested. The only real-thick-slice evaluation is qualitative, on 66 images with no paired ground truth. That's a load-bearing gap for the clinical promise, but not for the paper's core demonstration of a works-on-synthetic-degradation method.

Overall: this is a solid architectural contribution with a transparent evaluation. The generalization risk is real but the authors are honest about the need for future work on abnormal images. A good referee should push for error bars, a more realistic degradation model, and ideally some quantitative validation on real thick slices. The paper deserves a serious peer-review slot.

I'd bring this to a reading group; there's enough to discuss about synthetic degradation and GAN evaluation.

Referee Report

3 major / 4 minor

Summary. The paper proposes Virtual Thin Slice (VTS), a 3D conditional GAN for super-resolving thick-slice CT volumes (e.g., 8 mm interval) into virtual thin-slice volumes (1 mm interval). The generator is a 3D fully convolutional U-Net that predicts high-frequency residuals and can process arbitrary fields of view. The discriminator is conditioned on a vector encoding body part, slice interval, and Gaussian blur scale, with the stated goal of mitigating mode collapse. Training pairs are produced synthetically by a 'Degrader' that blurs, downsamples, and spline-interpolates thin-slice volumes. Experiments compare VTS against bicubic, SRCNN, and Pix2Pix on 53 test volumes using PSNR/SSIM and a Visual Turing Test with 8 participants; a qualitative evaluation on 66 real thick-slice CT volumes is also reported. The main claims are that VTS achieves the best quantitative and perceptual scores and can 'accurately reproduce' anatomy from real thick-slice CT.

Significance. If the results hold on real clinical data, the method could enable use of the vast archives of thick-slice CT for 3D visualization and analysis, which is a clinically valuable goal. The paper's strengths include a clearly described architecture, an ablation study showing the contribution of discriminator conditioning (Table 1: VTS PSNR 35.73 vs. 35.17 without conditioning), and the fully convolutional design that allows arbitrary input sizes and body parts with a single generator. The conditioning mechanism is simple and does not require extra information at test time. However, the quantitative validation is entirely on synthetically degraded data, and the real-data evaluation is qualitative only, so the central generalization claim is not yet supported by the evidence presented.

major comments (3)

[Section 3.3 and Section 4.2] The quantitative evaluation is performed exclusively on test volumes produced by the same synthetic Degrader model used for training (Gaussian blur with sigma up to 3.2 voxels, downsampling to 1/4 or 1/8, spline interpolation, and Gaussian noise). This measures how well the network inverts its own degradation model, not how it performs on real thick-slice CT acquisitions, which involve different slice sensitivity profiles, reconstruction kernels, helical interpolation, and quantum noise. The only real-data evaluation (Section 4.2, 66 images) is qualitative, with no paired real thick/thin volumes and no quantitative metrics. The abstract's claim of 'accurate reproduction of the principle anatomy' therefore goes beyond the evidence. Please add a quantitative validation on real paired thick/thin CT data, or substantially temper the clinical generalization claims.
[Table 1 and Visual Turing Test (Section 4.2)] The reported PSNR/SSIM differences are small: VTS achieves 35.73 dB/0.933 versus Pix2Pix's 35.14 dB/0.925 and the unconditioned variant's 35.17 dB/0.924. No error bars, standard deviations, or statistical significance tests are reported, so it is unclear whether the improvement over Pix2Pix or over the ablation is meaningful given typical run-to-run variation in GAN training. Similarly, the Visual Turing Test involved only 8 participants and 50 trials, and the 'roughly 90%' preference for VTS is presented without confidence intervals or a significance test. Please report per-volume statistics, error bars, and appropriate significance tests (e.g., paired tests on PSNR/SSIM and a binomial test on VTT preferences).
[Section 4.1 and Conclusion] The training data were 'carefully selected to not contain metal artifacts or noises because the discriminator is prone to reproduce such artifacts' (Section 4.1), and the paper acknowledges that 'in-depth evaluation on abnormal images is an important next step' (Conclusion). This is an explicit limitation that directly affects the clinical utility claimed in the abstract. Real clinical CT volumes frequently contain metal artifacts, noise, and pathologies. The current experiments do not demonstrate that the method handles such cases, and the qualitative statement that the model did not produce artifacts on some test data is not sufficient. Please clarify the intended scope of the claim and, if clinical utility is claimed, evaluate on a dataset that includes such cases.

minor comments (4)

[Related Work, first sentence] Typo: 'Single image super resolution is a major problems' should be 'a major problem'.
[Section 3.4] The conditioning vector is described as containing body part, slice interval (4 mm or 8 mm), and sigma scale (2 scales), totaling 8 channels. It would be clearer to state the exact encoding, e.g., one-hot vectors concatenated per channel, and how the 8 channels are distributed among the three factors.
[Figure 2 caption] The phrase 'we feed 1 mm spline interpolated 3D thick slice image itself' is ambiguous: it could mean the thick-slice volume is first resampled to 1 mm spacing before being input to the generator. Please state this preprocessing explicitly in the text (Section 3.3) as well.
[Section 4.2] The real thick-slice test covers slice intervals from 3.0 to 10.0 mm, while training simulates only 4 mm and 8 mm intervals. The paper does not explain how the network generalizes to 3 mm or 10 mm inputs; at minimum, this should be discussed as a distribution shift.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: VTS is trained and evaluated against external baselines on synthetically degraded thin-slice CT, and the core claims are not forced by construction.

full rationale

The paper's central derivation is a conditional GAN trained on paired thick/thin volumes, where thick volumes are produced by the Degrader (Gaussian blur, downsampling, spline interpolation, noise) from ground-truth thin volumes. This is a standard supervised super-resolution setup, not a circular one: the Degrader is a data-generation procedure, not a fitted parameter, and the network is tested on held-out thin volumes degraded in the same way. The discriminator conditioning vector (body part, slice interval, sigma) is an input to the model rather than a fitted target, and the ablation study (Table 1) compares VTS against an unconditioned variant, so the claimed benefit of conditioning has independent empirical content. All quantitative results are benchmarked against external baselines (bicubic, SRCNN, Pix2Pix) adapted to 3D, and the paper does not rely on self-citations or imported uniqueness theorems. The skeptical concern about whether the synthetic Degrader faithfully models real CT acquisition is a validity/generalization risk, not a circularity: the paper's claims are qualified by the synthetic test setup, and the real-thick-slice verification is explicitly qualitative. Therefore no circular step is present.

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

The central claim rests on the assumption that the synthetic degradation model (Gaussian blur + downsampling + spline) faithfully represents real thick-slice CT. Hyperparameters such as the L1 loss weight and conditioning vector design are hand-chosen and not specified fully, affecting reproducibility. No invented entities are introduced.

free parameters (3)

L1 loss weight lambda = not specified
Appears in Eq. (2) but its value is never given. It is a hand-chosen hyperparameter required to reproduce the method.
Conditioning vector composition = 8 channels: body part (4 classes), slice interval (4mm/8mm), sigma scale (2 levels)
The choice of what information to include and how to encode it (one-hot) is hand-designed and may affect performance.
Network capacity (initial channels, depth) = 64 channels, 4 downsampling stages, 4x4x4 kernels
Chosen to balance quality and computation; affects the central claim that the method works.

assumptions (4)

domain assumption The Degrader procedure (Gaussian blur with sigma in [0,3.2] voxels, downsample to 1/4 or 1/8, spline interpolation) accurately simulates the physical process of thick-slice CT acquisition from thin-slice CT.
This is the core training signal: the network learns to invert this degradation. If real thick-slice CT differs (noise, reconstruction, partial volume), the method may not generalize. Entered in Section 3.3.
domain assumption PSNR and SSIM are appropriate proxies for clinical image quality.
The paper uses these metrics as primary evidence of performance without validating them against diagnostic utility.
domain assumption The visual Turing test preferences of 8 radiology technicians/scientists reflect meaningful image quality differences.
Small sample size; no statistical significance reported, so the 90% preference claim may not be robust.
domain assumption Training data selection (excluding metal artifacts and noisy images) still yields a model representative of clinical CT data.
The paper acknowledges this selection in Section 4.1, yet claims to handle real thick CT images with artifacts.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Virtual Thin Slice: 3D Conditional GAN-based Super-resolution for CT Slice Interval." pith.science (2026). https://pith.science/paper/3A4IEL72

@misc{pith2026190811506,
  author       = {Pith},
  title        = {Pith review of: Virtual Thin Slice: 3D Conditional GAN-based Super-resolution for CT Slice Interval},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/3A4IEL72}},
  note         = {Machine review of arXiv:1908.11506}
}

read the original abstract

Many CT slice images are stored with large slice intervals to reduce storage size in clinical practice. This leads to low resolution perpendicular to the slice images (i.e., z-axis), which is insufficient for 3D visualization or image analysis. In this paper, we present a novel architecture based on conditional Generative Adversarial Networks (cGANs) with the goal of generating high resolution images of main body parts including head, chest, abdomen and legs. However, GANs are known to have a difficulty with generating a diversity of patterns due to a phenomena known as mode collapse. To overcome the lack of generated pattern variety, we propose to condition the discriminator on the different body parts. Furthermore, our generator networks are extended to be three dimensional fully convolutional neural networks, allowing for the generation of high resolution images from arbitrary fields of view. In our verification tests, we show that the proposed method obtains the best scores by PSNR/SSIM metrics and Visual Turing Test, allowing for accurate reproduction of the principle anatomy in high resolution. We expect that the proposed method contribute to effective utilization of the existing vast amounts of thick CT images stored in hospitals.

Figures

Figures reproduced from arXiv: 1908.11506 by the authors.

**Figure 1.** Comparison of original thick image and the virtual thin image output generated by the proposed approach. On the left top row, the CT sagittal view of original thick image is blurred with each vertebrae bone being nearly indistinguishable, while they become clear in corresponding ×8 super resolution image (Virtual Thin). Arbitrary size data, even the whole body shown on the right, is available for inputs and capable … view at source ↗

**Figure 2.** Adversarial training framework of thick-thin slice translation on CT images. In each training iteration, the thin slice input data is randomly degraded to simulate thick slice data. For the Generator input, we feed 1 mm spline interpolated 3D thick slice image itself. On the other hand, for the Discriminator input, we feed generated Virtual Thin image with slice condition including body part information with degrade… view at source ↗

**Figure 3.** Overview of the generator and discriminator network architectures. (x 3 × c) denotes the size of 3D feature map volumes, where c denotes the number of channels. Discriminator For the discriminator input, thick image (1 channel), thin image (1 channel) and slice information (8 channels) is given. Also a self-attention layer [15] is added in the fourth layer of the network. Self-attention mechanism is an idea to intr… view at source ↗

Figures from the paper (3 more)

**Figure 1.** Figure 1: The entire images are naturally reconstructed with no seams. [PITH_FULL_IMAGE:figures/full_fig_p007_1.png]

**Figure 4.** Figure 4: Comparison of the generated images of bicubic, SRCNN [3], pix2pix [7], VTS (ours) and corresponding ground truth thin slice image. [ 8× slice interpolation] [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]

**Figure 6.** Figure 6: Results using bicubic, SRCNN [3], pix2pix [7], VTS (ours) and corresponding ground truth thin slice image [PITH_FULL_IMAGE:figures/full_fig_p010_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

15 extracted references · 6 canonical work pages

[1]

In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition

Blau, Y., Michaeli, T.: The perception-distortion tradeoﬀ. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. pp. 6228–6237 (2018)

work page 2018
[2]

In: International Conference on Medical Image Computing and Computer-Assisted Intervention

Chen, Y., Shi, F., Christodoulou, A.G., Xie, Y., Zhou, Z., Li, D.: Eﬃcient and accurate mri super-resolution using a generative adversarial network and 3d multi- level densely connected network. In: International Conference on Medical Image Computing and Computer-Assisted Intervention. pp. 91–99. Springer (2018) Virtual Thin Slice 9

work page 2018
[3]

In: European conference on computer vision

Dong, C., Loy, C.C., He, K., Tang, X.: Learning a deep convolutional network for image super-resolution. In: European conference on computer vision. pp. 184–199. Springer (2014)

2014
[4]

Journal of applied meteorology 18(8), 1016–1022 (1979)

Duchon, C.E.: Lanczos ﬁltering in one and two dimensions. Journal of applied meteorology 18(8), 1016–1022 (1979)

work page 1979
[5]

In: Advances in neural information processing systems

Goodfellow, I., Pouget-Abadie, J., Mirza, M., Xu, B., Warde-Farley, D., Ozair, S., Courville, A., Bengio, Y.: Generative adversarial nets. In: Advances in neural information processing systems. pp. 2672–2680 (2014)

2014
[6]

arXiv preprint arXiv:1502.03167 (2015)

Ioﬀe, S., Szegedy, C.: Batch normalization: Accelerating deep network training by reducing internal covariate shift. arXiv preprint arXiv:1502.03167 (2015)

arXiv 2015
[7]

arXiv preprint (2017)

Isola, P., Zhu, J.Y., Zhou, T., Efros, A.A.: Image-to-image translation with condi- tional adversarial networks. arXiv preprint (2017)

work page 2017
[8]

arXiv preprint (2017)

Ledig, C., Theis, L., Husz´ ar, F., Caballero, J., Cunningham, A., Acosta, A., Aitken, A., Tejani, A., Totz, J., Wang, Z., et al.: Photo-realistic single image super- resolution using a generative adversarial network. arXiv preprint (2017)

work page 2017

Show all 15 references

[9]

arXiv preprint arXiv:1611.02163 (2016)

Metz, L., Poole, B., Pfau, D., Sohl-Dickstein, J.: Unrolled generative adversarial networks. arXiv preprint arXiv:1611.02163 (2016)

2016 arXiv
[10]

arXiv preprint arXiv:1411.1784 (2014)

Mirza, M., Osindero, S.: Conditional generative adversarial nets. arXiv preprint arXiv:1411.1784 (2014)

2014 arXiv
[11]

arXiv preprint arXiv:1511.06434 (2015)

Radford, A., Metz, L., Chintala, S.: Unsupervised representation learning with deep convolutional generative adversarial networks. arXiv preprint arXiv:1511.06434 (2015)

2015 arXiv
[12]

In: International Conference on Medical image computing and computer-assisted intervention

Ronneberger, O., Fischer, P., Brox, T.: U-net: Convolutional networks for biomedi- cal image segmentation. In: International Conference on Medical image computing and computer-assisted intervention. pp. 234–241. Springer (2015)

2015
[13]

IEEE transactions on image processing 13(4), 600–612 (2004)

Wang, Z., Bovik, A.C., Sheikh, H.R., Simoncelli, E.P.: Image quality assessment: from error visibility to structural similarity. IEEE transactions on image processing 13(4), 600–612 (2004)

2004
[14]

In: European Conference on Computer Vision

Yang, C.Y., Ma, C., Yang, M.H.: Single-image super-resolution: A benchmark. In: European Conference on Computer Vision. pp. 372–386. Springer (2014)

2014
[15]

Zhang, H., Goodfellow, I., Metaxas, D., Odena, A.: Self-attention generative ad- versarial networks. arXiv preprint arXiv:1805.08318 (2018) Acknowledgements We acknowledge using the Reedbush-L (SGI Rackable C2112-4GP3/C1102-GP8) HPC system in the Information Technology Center,...

2018 arXiv

Pith tools

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

[1] [1]

In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition

Blau, Y., Michaeli, T.: The perception-distortion tradeoﬀ. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. pp. 6228–6237 (2018)

work page 2018

[2] [2]

In: International Conference on Medical Image Computing and Computer-Assisted Intervention

Chen, Y., Shi, F., Christodoulou, A.G., Xie, Y., Zhou, Z., Li, D.: Eﬃcient and accurate mri super-resolution using a generative adversarial network and 3d multi- level densely connected network. In: International Conference on Medical Image Computing and Computer-Assisted Intervention. pp. 91–99. Springer (2018) Virtual Thin Slice 9

work page 2018

[3] [3]

In: European conference on computer vision

Dong, C., Loy, C.C., He, K., Tang, X.: Learning a deep convolutional network for image super-resolution. In: European conference on computer vision. pp. 184–199. Springer (2014)

2014

[4] [4]

Journal of applied meteorology 18(8), 1016–1022 (1979)

Duchon, C.E.: Lanczos ﬁltering in one and two dimensions. Journal of applied meteorology 18(8), 1016–1022 (1979)

work page 1979

[5] [5]

In: Advances in neural information processing systems

Goodfellow, I., Pouget-Abadie, J., Mirza, M., Xu, B., Warde-Farley, D., Ozair, S., Courville, A., Bengio, Y.: Generative adversarial nets. In: Advances in neural information processing systems. pp. 2672–2680 (2014)

2014

[6] [6]

arXiv preprint arXiv:1502.03167 (2015)

Ioﬀe, S., Szegedy, C.: Batch normalization: Accelerating deep network training by reducing internal covariate shift. arXiv preprint arXiv:1502.03167 (2015)

arXiv 2015

[7] [7]

arXiv preprint (2017)

Isola, P., Zhu, J.Y., Zhou, T., Efros, A.A.: Image-to-image translation with condi- tional adversarial networks. arXiv preprint (2017)

work page 2017

[8] [8]

arXiv preprint (2017)

Ledig, C., Theis, L., Husz´ ar, F., Caballero, J., Cunningham, A., Acosta, A., Aitken, A., Tejani, A., Totz, J., Wang, Z., et al.: Photo-realistic single image super- resolution using a generative adversarial network. arXiv preprint (2017)

work page 2017