Combined Image Data Augmentations diminish the benefits of Adaptive Label Smoothing

Andrey Morozov; Ekagra Gupta; Georg Siedel; Silvia Vock; Weijia Shao

REVIEW 4 major objections 6 minor 40 references

Combined Image Data Augmentations diminish the benefits of Adaptive Label Smoothing

T0 review · 4 major / 6 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read Adaptive label smoothing helps when one augmentation dominates, but vanishes under diverse transforms.

desk verdict Useful empirical mapping of where adaptive label smoothing helps, but the 'benefits vanish' claim is underpowered and should be softened. read the letter →

arxiv 2507.16427 v1 pith:F2FRCDNO submitted 2025-07-22 cs.CV cs.LG

Georg Siedel , Ekagra Gupta , Weijia Shao , Silvia Vock , Andrey Morozov This is my paper · ORCID

classification cs.CVcs.LG

keywords adaptivelabelsmoothingsoftaugmentationTrivialAugmentRandomErasingnoiseinjectioncorruptionrobustnessimageclassificationdata

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

The paper extends adaptive label smoothing—softening a training image's label in proportion to how much an augmentation degrades it—from random crops to Random Erasing, noise injection, and TrivialAugment. It shows that soft Random Erasing allows more aggressive erasing without hurting accuracy or robustness, and soft noise helps marginally. The central negative result is that when a wide, heterogeneous set of transformations is applied together as in TrivialAugment, the gains from adaptive smoothing disappear, and aggressive smoothing actually reduces robustness to common corruptions. The authors conclude that adaptive label smoothing should be reserved for training pipelines dominated by one homogeneous transformation type.

What carries the argument

The central object is the magnitude-to-confidence mapping, a function $\alpha(\phi)$ that translates an augmentation's severity—rotation angle, occluded area ratio, contrast adjustment, noise intensity—into the smoothed label target $g_{\alpha}(y)$. The paper builds these mappings from human vision studies reporting classification accuracy under distortion, from the accuracy of a pretrained proxy model on transformed training images, from image-similarity metrics (SSIM, NCC, SCC, UIQ, SIFT), and from a polynomial fallback $\alpha = \phi^k (1 - p_{\min})$ with parameters fit to the human-vision curve or to model accuracy. The framework couples input and label perturbations: a transformed image receives a label whose confidence falls with the transformation magnitude. The machinery's job is to test whether the random-crop success of this coupling carries over; empirically it does for single-transform schemes such as Random Erasing, but not for TrivialAugment, where the diverse transformation set appears to provide the regularization itself.

What would settle it

Train on CIFAR-100 and TinyImageNet with TrivialAugment while softening labels with per-transformation curves measured directly on the exact transformed images (for example, human accuracy on those images at the precise TrivialAugment magnitudes), instead of curves transferred between transformation types. If this condition-specific soft TrivialAugment matches or beats hard-label TrivialAugment in accuracy or corruption robustness by more than one standard deviation over five runs, the paper's claim that the benefit vanishes under diverse transform sets would be falsified.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that magnitude-conditioned label confidence transfers from random crops to single-transform augmentations but not to heterogeneous mixtures. Soft Random Erasing lowers the label confidence in proportion to the occluded area ratio, which allows the erasing probability and maximum area ratio to be raised well beyond what hard-label training tolerates, improving CIFAR accuracy. Soft Gaussian and Patch Gaussian noise give smaller, less consistent gains, and soft random cropping remains the strongest accuracy booster while consistently hurting corruption robustness. For TrivialAugment, which randomly picks one of fourteen transformations and a magnitude, none of the tested mappings—human-vision-based, proxy-model-accuracy-based, polynomial, or image-similarity-based—produces a meaningful gain; aggressive label smoothing in this setting degrades both accuracy and robustness to common corruptions. Isolating each transformation shows that softening helps individual transforms, but the benefit disappears as transformations are accumulated, and softening only the three most beneficial ones still fails to beat standard TrivialAugment.

Load-bearing premise

The load-bearing premise is that the magnitude-to-confidence curves—taken from human-vision studies (including transfers from rotation to shear and contrast to brightness), a proxy model's accuracy on transformed images, and the chosen polynomial fallbacks—faithfully measure how much information a transformation removes for the dataset at hand; if those curves are miscalibrated, the null result under TrivialAugment could be an artifact of the mapping rather than an intrinsic limit of adaptive smoothing.

Editorial extensions

If this is right

Random Erasing pipelines can be pushed to higher application probabilities and larger occlusion areas when labels are softened in proportion to the occluded area, improving accuracy over hard-label training.
For TrivialAugment and similar multi-transform schemes, adaptive label smoothing adds little; diversity of transformation types already supplies the regularization, so effort is better spent on tuning the augmentation policy itself.
Aggressive label smoothing on a heterogeneous augmentation set reduces robustness to common corruptions, so any remaining smoothing should be clamped to conservative magnitudes.
Because several soft-label gains fall within run-to-run standard deviation, adopting adaptive smoothing for a new dataset should be validated empirically rather than assumed beneficial.

Reading between the lines

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

The isolated-per-transform gains suggest an untested schedule in which smoothing is applied only while a single augmentation dominates and then turned off as diversity increases; this could recover part of the benefit in mixed pipelines.
If the magnitude-to-confidence curves are miscalibrated, the negative TrivialAugment result might reflect curve error rather than a true ceiling; image-content-conditioned smoothing, which the authors list as future work, would test this directly.
The robustness drop from aggressive smoothing hints that label confidence should be tied to whether the transformation removes class-relevant information per sample, not just to its average severity.
The transfer practice used in the paper—borrowing a rotation study for shear and a contrast study for brightness—could be checked against direct measurements of those transformations, providing a cheap validity test of the mapping framework.

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.

Desk Editor's Note

read the letter

The paper extends soft augmentation beyond random crop to random erasing, noise injections, and TrivialAugment, and it does this carefully. The mapping analysis — comparing human vision data, proxy model accuracy, and image similarity metrics — is thoughtful, and dropping the image metrics because they are inconsistent is the right call. The negative result for TrivialAugment is genuinely new: under a heterogeneous augmentation policy, adaptive label smoothing gives no clear gain. That is useful guidance for practitioners, and the experiments span three datasets and two architectures. The writing is clear, and the authors are honest that many differences fall within standard deviations.

The main soft spot is the central "benefits vanish" claim. It is a zero-effect claim, but it rests on 5 runs per configuration. Table 2 actually shows positive point estimates for the mild k=2, pmin=0.7 mapping on CIFAR-10 (96.78 vs 96.62) and CIFAR-100 (80.37 vs 80.20), and fixed label smoothing improves TinyImageNet robustness. With n=5 and standard deviations around 0.1–0.2, a true gain of 0.1–0.2 accuracy points would almost certainly be missed. The authors hedge in the text ("little statistical significance"), but the abstract and conclusion state the negative result flatly. That overstates the evidence. The fix is straightforward: more runs, or a proper Bayesian treatment, and language that says "no significant benefit detected" rather than "benefits vanish."

Other issues are real but secondary. RE hyperparameters were tuned on CIFAR-100 and then used to report CIFAR-100 results, which is selection bias. The code link in the abstract is malformed (spaces in the URL). The mapping transfers — rotation HVS data for shear, contrast for brightness — are plausible but unvalidated; if those curves are miscalibrated, the soft TA comparison is not a fair test. That is a legitimate limitation, and the authors acknowledge it, but it should be discussed more prominently.

Overall, the core contribution holds up: this is a solid empirical study that maps out the regime where adaptive label smoothing helps, and it provides a useful warning about combining it with heterogeneous augmentations. The statistical concerns are addressable and do not invalidate the main finding. The paper deserves a serious referee; I would send it to review with a request for more runs, a corrected code link, and a more measured abstract.

Who should read it: anyone choosing augmentation pipelines or working on label smoothing. It is not a breakthrough, but it is a careful, useful data point.

Referee Report

4 major / 6 minor

Summary. This paper extends adaptive label smoothing ('soft augmentation') beyond random cropping to Random Erasing, Gaussian/Patch-Gaussian noise, and the TrivialAugment policy. The authors construct magnitude-to-confidence mappings from human vision studies, proxy-model accuracy curves, and image similarity metrics, then evaluate WideResNet and ResNeXt models on CIFAR-10/100 and TinyImageNet. They report that soft labels allow more aggressive Random Erasing settings, give mixed results for noise injection, and yield no significant benefit when applied within TrivialAugment or when combined with multiple augmentations, often reducing robustness to common corruptions. The paper concludes that adaptive label smoothing is only useful when a single homogeneous augmentation type dominates the training pipeline.

Significance. If the negative result for TrivialAugment is statistically robust, the paper usefully delimits the applicability of adaptive label smoothing and provides a reproducible starting point for future work on augmentation-aware label confidence. The authors test several mapping strategies, report standard deviations, and release code, which are strengths. However, the central 'benefits vanish' claim rests on five-run comparisons that are underpowered for detecting small effects, and the positive Random Erasing claim is inconsistent across datasets and lacks essential hyperparameter reporting. The significance of the paper therefore depends on whether the authors can strengthen the statistical evidence and clarify the scope of their conclusions.

major comments (4)

[§4.1, Table 2, Figure 4] The central claim that the benefits of adaptive label smoothing 'vanish' under TrivialAugment is an inference from the absence of a significant improvement in five-run comparisons. The paper's own caption acknowledges that no improvement exceeds the combined standard deviations, but a failure to reject the null at n=5 does not establish zero effect; the k=2, pmin=0.7 mapping shows positive point estimates of +0.16 on CIFAR-10 and +0.17 on CIFAR-100 accuracy. Please report confidence intervals or an equivalence test, and revise the abstract and conclusion to state that no significant or practically meaningful benefit was detected rather than that the benefit vanishes.
[§4.1] The statement that 'Even when we use all transformations, but apply adaptive label smoothing only to the 3 transformation types that benefit the most from it, the model accuracy is no better than for standard TA' is not backed by any table, figure, or specification of which three transformation types were selected. This experiment is part of the argument that the benefit disappears under diversity, so it should be reported with the same level of detail as the other results, including which transforms were used and the resulting accuracies.
[§4.2, Table 3] The positive claim that soft Random Erasing enables stronger regularization is not reproducible because the hyperparameters selected in the sweep (application probability and maximum area ratio) are never reported. In addition, soft RE on TinyImageNet underperforms standard RE (60.05 vs 60.52), an effect attributed to reusing CIFAR parameters without providing a supporting experiment with TinyImageNet-tuned parameters. Please report the selected RE settings and either tune per dataset or temper the claim accordingly.
[§3.2, Table 1] The magnitude-to-confidence curves for ShearX/ShearY are transferred from rotation HVS data and for Brightness from contrast HVS data, with the stated justification that proxy-model behavior and image similarity metrics behave similarly. This transfer is an assumption that could bias the soft-TA comparison if the proxy curves are miscalibrated on the target datasets. Because the negative result is the paper's main contribution, please add a sensitivity check with at least one alternative mapping for the transferred transforms, or provide the quantitative similarity evidence referenced in the text.

minor comments (6)

[Throughout] There are formatting typos in the captions, including 'T able 1', 'T able 2', and 'CIF AR-10'; these should be fixed before publication.
[Table 4] Several entries in Table 4 are run together without separators, for example '+0.33 ±0.10 -0.14±0.97+0.42±0.35'; the table should be reformatted so that each cell contains a single mean and standard deviation.
[§3.1, Fig. 2] The polynomial mapping is presented as α(ϕ)=ϕk(1−pmin) where pmin is called the 'minimum confidence', yet the figures plot 'Label Confidence' on the y-axis. If α is the smoothing factor, then the plotted confidence is 1−α; the manuscript should define both quantities explicitly to avoid ambiguity about whether the mapping lowers confidence or raises the smoothing fraction.
[Figure 4 caption] The caption uses 'k = 2, p≥ chance' which should read 'k = 2, pmin = chance' for consistency with the notation in Section 3.1.
[Appendix A.1] The line 'Random Erasing value Random standard gaussian' is unclear; please specify the exact distribution and parameters used for the erasing fill value.
[§2.3] The reweighting mechanism is cited to [21] but no explicit loss formula is given; since reweighted variants appear in Tables 2 and 3, the exact reweighting term should be defined in the main text or appendix.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: label-confidence mappings come from external HVS data and an independent proxy model, and the main comparisons are not forced by construction.

full rationale

The paper's derivation chain is empirical rather than definitional. The magnitude-to-confidence mappings are fixed before the main experiments from independent sources: HVS accuracy studies (refs 7, 11-12, 27), an external proxy classifier pretrained with Mixup and noise injections but 'with no transformations contained in TA or RE' (Section 3.1), and image similarity metrics. These sources do not use the final test accuracies of the WRN/ResNeXt models, so the soft-TA accuracy and robustness results are not determined by construction. The k=2, pmin=chance mapping is explicitly inherited from prior work (Liu et al. [16]) as a baseline, and k=2, pmin=0.7 is anchored to the standard 0.1 average smoothing factor rather than fitted to the outcome. The Random Erasing sweep (Figure 5) directly compares soft versus hard labels across the same hyperparameter grid, and the noise results in Table 3 are comparisons against hard-label counterparts. The one same-author citation ([25]) appears only in related work about p-norm noise and is not load-bearing. The paper's own caveat that 'no improvement over the baseline is larger than the combined standard deviations' is a statistical-power limitation, not a circular reduction; likewise, selecting the most promising mapping on the same benchmarks is selection bias but does not make the central claim true by definition. No circular step can be quoted because no fitted parameter is renamed as a prediction and no result is forced by self-citation.

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

The central claims rest on several manually chosen constants and domain assumptions about how well external human-vision data and proxy model accuracy capture information loss. No new physical or conceptual entities are introduced.

free parameters (3)

Polynomial mapping curvature k and minimum confidence pmin = k=2; pmin in {0.7, chance, 0.3} depending on scheme
Chosen by hand per mapping scheme; these constants determine how strongly label confidence is reduced as a function of augmentation magnitude.
k_est and pmin_est for the k = est. mapping = values not reported in text; fit to HVS or proxy model accuracy curves
Fit to external HVS/model-accuracy data to approximate the curve, not to the target classification accuracy, but still free parameters of the method.
Random Erasing application probability and maximum area ratio = specific values not stated in text; optimised via grid sweep in Fig. 5 on CIFAR-100 and reused for TIN
Swept and optimised separately for standard and soft RE on CIFAR-100 before final comparisons, introducing selection.

assumptions (3)

domain assumption Human vision accuracy data measured for one transformation type transfers to other transformation types (rotation to shear, contrast to brightness).
Invoked in Section 3.2 and Table 1 without direct validation, based on observed similarity of proxy model behavior and image similarity metrics.
domain assumption A proxy model's accuracy on transformed training images quantifies information loss and therefore appropriate label confidence.
Used in Section 3.1 model-accuracy mapping; assumes the chosen pretrained proxy model is a valid judge of distortion severity across all transformations.
domain assumption Label confidence should drop uniformly across all images, classes, and datasets for a given transformation magnitude.
Explicitly acknowledged in Section 5 as a simplifying assumption; the paper does not condition smoothing on image content or class.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Combined Image Data Augmentations diminish the benefits of Adaptive Label Smoothing." pith.science (2026). https://pith.science/paper/F2FRCDNO

@misc{pith2026250716427,
  author       = {Pith},
  title        = {Pith review of: Combined Image Data Augmentations diminish the benefits of Adaptive Label Smoothing},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/F2FRCDNO}},
  note         = {Machine review of arXiv:2507.16427}
}

read the original abstract

Soft augmentation regularizes the supervised learning process of image classifiers by reducing label confidence of a training sample based on the magnitude of random-crop augmentation applied to it. This paper extends this adaptive label smoothing framework to other types of aggressive augmentations beyond random-crop. Specifically, we demonstrate the effectiveness of the method for random erasing and noise injection data augmentation. Adaptive label smoothing permits stronger regularization via higher-intensity Random Erasing. However, its benefits vanish when applied with a diverse range of image transformations as in the state-of-the-art TrivialAugment method, and excessive label smoothing harms robustness to common corruptions. Our findings suggest that adaptive label smoothing should only be applied when the training data distribution is dominated by a limited, homogeneous set of image transformation types.

Figures

Figures reproduced from arXiv: 2507.16427 by the authors.

**Figure 1.** TinyImageNet images transformed with soft TrivialAugment and soft Random Erasing (RE) data augmentation. The titles display the TrivialAugment transformation type and whether RE is applied, as well as the label and its softened confidence. The confidence is calculated as a function of the augmentation severity for every transformation type individually. Here, the functions are derived from a proxy models average ac… view at source ↗

**Figure 2.** This diagram displays the functions defining how an images label is adjusted based on the magnitude of the image transformation. For the ”Rotate” transformation in TrivialAugment, the functions map label confidence to image rotation in degrees. The functions are based on human vision studies (blue), on a proxy models outputs (green), on custom polynomial functions (red) and on image similarity metrics (black). Here,… view at source ↗

**Figure 3.** A summary of the functions defining how an images label is adjusted based on the magnitude of the image transformation, for all transformation types used in TrivialAugment (find a closer view for one transformation type in [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗

Figures from the paper (2 more)

**Figure 4.** Figure 4: Standard TrivialAugment and TrivialAugment with adaptive label smoothing (soft) are compared for individual transformation types from its set (left) and when incrementally adding transformations (right), evaluating accuracy (top) and robustness (bottom). The transforma…

**Figure 5.** Figure 5: Parameter sweep of the two main hyperparameters of Random Erasing on CIFAR-100. Displayed are the mean and standard deviations for accuracy over 5 runs for standard and soft Random Erasing. No reweighting or other augmentations are applied. that adaptive label smoothin…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

40 extracted references · 27 canonical work pages

[1]

Journal of neurophysiology 87(6), 3102–3116 (2002)

Avidan, G., Harel, M., Hendler, T., Ben-Bashat, D., Zohary, E., Malach, R.: Con- trast sensitivity in human visual areas and its relationship to object recognition. Journal of neurophysiology 87(6), 3102–3116 (2002)

work page 2002
[2]

arXiv preprint arXiv:1805.02641 (2018)

Bagherinezhad, H., Horton, M., Rastegari, M., Farhadi, A.: Label refinery: Improving imagenet classification through label progression. arXiv preprint arXiv:1805.02641 (2018)

arXiv 2018
[3]

In: 13th British Machine Vision Conference

Brown, M., Lowe, D.: Invariant features from interest point groups. In: 13th British Machine Vision Conference. vol. 4, pp. 398–410 (2002)

work page 2002
[4]

In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition

Cubuk, E.D., Zoph, B., Mane, D., Vasudevan, V., Le, Q.V.: Autoaugment: Learning augmentation strategies from data. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. pp. 113–123 (2019)

work page 2019
[5]

In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition workshops

Cubuk, E.D., Zoph, B., Shlens, J., Le, Q.V.: Randaugment: Practical automated data augmentation with a reduced search space. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition workshops. pp. 702–703 (2020)

2020
[6]

arXiv preprint arXiv:1708.04552 (2017)

DeVries, T., Taylor, G.W.: Improved regularization of convolutional neural net- works with cutout. arXiv preprint arXiv:1708.04552 (2017)

arXiv 2017
[7]

ACM Transactions on Applied Perception (TAP) 16(2), 1–17 (2019)

Dodge, S., Karam, L.: Human and dnn classification performance on images with quality distortions: A comparative study. ACM Transactions on Applied Perception (TAP) 16(2), 1–17 (2019)

work page 2019
[8]

In: Proceedings of The 27th International Conference on Artificial Intelligence and Statistics

Erichson, B., Lim, S.H., Xu, W., Utrera, F., Cao, Z., Mahoney, M.: NoisyMix: Boosting model robustness to common corruptions. In: Proceedings of The 27th International Conference on Artificial Intelligence and Statistics. Proceedings of Machine Learning Research, vol. 238, pp. 4033–4041. PMLR (02–04 May 2024)

work page 2024

Show all 40 references

[9]

Neuron 47(4), 607– 620 (2005)

Gardner, J.L., Sun, P., Waggoner, R.A., Ueno, K., Tanaka, K., Cheng, K.: Contrast adaptation and representation in human early visual cortex. Neuron 47(4), 607– 620 (2005)

2005
[10]

International Conference on Learning Representa- tions 2019 p

Hendrycks, D., Dietterich, T.: Benchmarking neural network robustness to common corruptions and perturbations. International Conference on Learning Representa- tions 2019 p. 16 (28032019)

2019
[11]

Science 218(4574), 804–806 (1982)

Hollard, V.D., Delius, J.D.: Rotational invariance in visual pattern recognition by pigeons and humans. Science 218(4574), 804–806 (1982)

1982
[12]

Journal of Ex- perimental Child Psychology 40(1), 181–192 (1985)

Kail, R.: Development of mental rotation: A speed-accuracy study. Journal of Ex- perimental Child Psychology 40(1), 181–192 (1985)

1985
[13]

Krizhevsky, A., Hinton, G., et al.: Learning multiple layers of features from tiny images (2009)

2009
[14]

Advances in neural information processing systems 25 (2012)

Krizhevsky, A., Sutskever, I., Hinton, G.E.: Imagenet classification with deep con- volutional neural networks. Advances in neural information processing systems 25 (2012)

2012
[15]

CS 231N, 2015 (2015) 14 Siedel et al

Le, Y., Yang, X.: Tiny imagenet visual recognition challenge. CS 231N, 2015 (2015) 14 Siedel et al

2015
[16]

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

Liu, Y., Yan, S., Leal-Taix´ e, L., Hays, J., Ramanan, D.: Soft augmentation for image classification. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 16241–16250 (2023)

2023
[17]

arXiv preprint arXiv:1906.02611 (2019)

Lopes, R.G., Yin, D., Poole, B., Gilmer, J., Cubuk, E.D.: Improving robustness without sacrificing accuracy with patch gaussian augmentation. arXiv preprint arXiv:1906.02611 (2019)

2019 arXiv
[18]

Interna- tional journal of computer vision 60, 91–110 (2004)

Lowe, D.G.: Distinctive image features from scale-invariant keypoints. Interna- tional journal of computer vision 60, 91–110 (2004)

2004
[19]

Cell reports 40(13) (2022)

Mazade, R., Jin, J., Rahimi-Nasrabadi, H., Najafian, S., Pons, C., Alonso, J.M.: Cortical mechanisms of visual brightness. Cell reports 40(13) (2022)

2022
[20]

In: Proceedings of the IEEE/CVF international conference on computer vision

M¨ uller, S.G., Hutter, F.: Trivialaugment: Tuning-free yet state-of-the-art data aug- mentation. In: Proceedings of the IEEE/CVF international conference on computer vision. pp. 774–782 (2021)

2021
[21]

In: International conference on machine learning

Ren, M., Zeng, W., Yang, B., Urtasun, R.: Learning to reweight examples for robust deep learning. In: International conference on machine learning. pp. 4334–
[22]

In: European Conference on Computer Vision (ECCV)

Rusak, E., Schott, L., Zimmermann, R.S., Bitterwolf, J., Bringmann, O., Bethge, M., Brendel, W.: A simple way to make neural networks robust against diverse image corruptions. In: European Conference on Computer Vision (ECCV). pp. 53–69. Springer (2020)

2020
[23]

Proceedings of the national academy of sciences 104(15), 6424– 6429 (2007)

Serre, T., Oliva, A., Poggio, T.: A feedforward architecture accounts for rapid categorization. Proceedings of the national academy of sciences 104(15), 6424– 6429 (2007)

2007
[24]

Journal of big data 6(1), 1–48 (2019)

Shorten, C., Khoshgoftaar, T.M.: A survey on image data augmentation for deep learning. Journal of big data 6(1), 1–48 (2019)

2019
[25]

In: Proceedings of the 19th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications - Volume 2: VISAPP

Siedel, G., Shao, W., Vock, S., Morozov, A.: Investigating the corruption robust- ness of image classifiers with random p-norm corruptions. In: Proceedings of the 19th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications - Vo...

2024
[26]

In: Proceedings of the IEEE conference on computer vision and pattern recognition

Szegedy, C., Vanhoucke, V., Ioffe, S., Shlens, J., Wojna, Z.: Rethinking the incep- tion architecture for computer vision. In: Proceedings of the IEEE conference on computer vision and pattern recognition. pp. 2818–2826 (2016)

2016
[27]

Proceedings of the National Academy of Sciences 115(35), 8835–8840 (2018)

Tang, H., Schrimpf, M., Lotter, W., Moerman, C., Paredes, A., Ortega Caro, J., Hardesty, W., Cox, D., Kreiman, G.: Recurrent computations for visual pattern completion. Proceedings of the National Academy of Sciences 115(35), 8835–8840 (2018)

2018
[28]

Vryniotis, V.: How to train state-of-the-art models using torchvision’s latest primi- tives (2021), https://pytorch.org/blog/how-to-train-state-of-the-art-models-using- torchvision-latest-primitives/, 17.10.2023

2021
[29]

arXiv preprint arXiv:2009.09496 (2020)

Vyas, N., Saxena, S., Voice, T.: Learning soft labels via meta learning. arXiv preprint arXiv:2009.09496 (2020)

2020 arXiv
[30]

IEEE signal processing letters 9(3), 81–84 (2002)

Wang, Z., Bovik, A.C.: A universal image quality index. IEEE signal processing letters 9(3), 81–84 (2002)

2002
[31]

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
[32]

In: Proceedings of the IEEE conference on computer vision and pattern recognition

Xie, L., Wang, J., Wei, Z., Wang, M., Tian, Q.: Disturblabel: Regularizing cnn on the loss layer. In: Proceedings of the IEEE conference on computer vision and pattern recognition. pp. 4753–4762 (2016) Aggressive Soft Augmentation 15

2016
[33]

In: Proceedings of the IEEE conference on computer vision and pattern recognition

Xie, S., Girshick, R., Doll´ ar, P., Tu, Z., He, K.: Aggregated residual transformations for deep neural networks. In: Proceedings of the IEEE conference on computer vision and pattern recognition. pp. 1492–1500 (2017)

2017
[34]

Circuits, systems and sig- nal processing 28, 819–843 (2009)

Yoo, J.C., Han, T.H.: Fast normalized cross-correlation. Circuits, systems and sig- nal processing 28, 819–843 (2009)

2009
[35]

In: Proceedings of the IEEE/CVF international conference on computer vision

Yun, S., Han, D., Oh, S.J., Chun, S., Choe, J., Yoo, Y.: Cutmix: Regularization strategy to train strong classifiers with localizable features. In: Proceedings of the IEEE/CVF international conference on computer vision. pp. 6023–6032 (2019)

2019
[36]

In: British Machine Vision Conference 2016

Zagoruyko, S., Komodakis, N.: Wide residual networks. In: British Machine Vision Conference 2016. British Machine Vision Association (2016)

2016
[37]

IEEE Transactions on Image Processing 30, 5984–5996 (2021)

Zhang, C.B., Jiang, P.T., Hou, Q., Wei, Y., Han, Q., Li, Z., Cheng, M.M.: Delving deep into label smoothing. IEEE Transactions on Image Processing 30, 5984–5996 (2021)

2021
[38]

In: International Conference on Learning Representations (2018)

Zhang, H., Cisse, M., Dauphin, Y.N., Lopez-Paz, D.: mixup: Beyond empirical risk minimization. In: International Conference on Learning Representations (2018)

2018
[39]

In: Proceedings of the AAAI conference on artificial intelligence

Zhong, Z., Zheng, L., Kang, G., Li, S., Yang, Y.: Random erasing data augmenta- tion. In: Proceedings of the AAAI conference on artificial intelligence. vol. 34, pp. 13001–13008 (2020)

2020
[40]

In: Proceedings of the Annual Meeting of the Cognitive Science Society

Zhu, H., Tang, P., Park, J., Park, S., Yuille, A.: Robustness of object recognition under extreme occlusion in humans and computational models. In: Proceedings of the Annual Meeting of the Cognitive Science Society. vol. 41 (2019) 16 Siedel et al. A Appendix A.1 T raining para...

2019

Pith tools

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

[1] [1]

Journal of neurophysiology 87(6), 3102–3116 (2002)

Avidan, G., Harel, M., Hendler, T., Ben-Bashat, D., Zohary, E., Malach, R.: Con- trast sensitivity in human visual areas and its relationship to object recognition. Journal of neurophysiology 87(6), 3102–3116 (2002)

work page 2002

[2] [2]

arXiv preprint arXiv:1805.02641 (2018)

Bagherinezhad, H., Horton, M., Rastegari, M., Farhadi, A.: Label refinery: Improving imagenet classification through label progression. arXiv preprint arXiv:1805.02641 (2018)

arXiv 2018

[3] [3]

In: 13th British Machine Vision Conference

Brown, M., Lowe, D.: Invariant features from interest point groups. In: 13th British Machine Vision Conference. vol. 4, pp. 398–410 (2002)

work page 2002

[4] [4]

In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition

Cubuk, E.D., Zoph, B., Mane, D., Vasudevan, V., Le, Q.V.: Autoaugment: Learning augmentation strategies from data. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. pp. 113–123 (2019)

work page 2019

[5] [5]

In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition workshops

Cubuk, E.D., Zoph, B., Shlens, J., Le, Q.V.: Randaugment: Practical automated data augmentation with a reduced search space. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition workshops. pp. 702–703 (2020)

2020

[6] [6]

arXiv preprint arXiv:1708.04552 (2017)

DeVries, T., Taylor, G.W.: Improved regularization of convolutional neural net- works with cutout. arXiv preprint arXiv:1708.04552 (2017)

arXiv 2017

[7] [7]

ACM Transactions on Applied Perception (TAP) 16(2), 1–17 (2019)

Dodge, S., Karam, L.: Human and dnn classification performance on images with quality distortions: A comparative study. ACM Transactions on Applied Perception (TAP) 16(2), 1–17 (2019)

work page 2019

[8] [8]

In: Proceedings of The 27th International Conference on Artificial Intelligence and Statistics

Erichson, B., Lim, S.H., Xu, W., Utrera, F., Cao, Z., Mahoney, M.: NoisyMix: Boosting model robustness to common corruptions. In: Proceedings of The 27th International Conference on Artificial Intelligence and Statistics. Proceedings of Machine Learning Research, vol. 238, pp. 4033–4041. PMLR (02–04 May 2024)

work page 2024