mlsanityData to 7 Oct 2026

Stylized Synthetic Augmentation further improves Corruption Robustness

arXiv 2512.15675 ↗First posted 17 Dec 2025cs.CV cs.LGcited by 0

Not on the map. Why?

Abstract (arXiv)

This paper proposes a training data augmentation pipeline that combines synthetic image data with neural style transfer in order to address the vulnerability of deep vision models to common corruptions. We show that although applying style transfer on synthetic images degrades their quality with respect to the common Frechet Inception Distance (FID) metric, these images are surprisingly beneficial for model training. We conduct a systematic empirical analysis of the effects of both augmentations and their key hyperparameters on the performance of image classifiers. Our results demonstrate that stylization and synthetic data complement each other well and can be combined with popular rule-based data augmentation techniques such as TrivialAugment, while not working with others. Our method achieves state-of-the-art corruption robustness on several small-scale image classification benchmarks, reaching 93.54%, 74.9% and 50.86% robust accuracy on CIFAR-10-C, CIFAR-100-C and TinyImageNet-C, respectively

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 52 of its 82 references that are in the map. First posted on arXiv 17 Dec 2025.

Builds on

52 of 82 references.

References · 52

  1. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 29,915
  2. 2020Denoising Diffusion Probabilistic Modelscited by 20,851
  3. 2021High-Resolution Image Synthesis with Latent Diffusion Modelscited by 18,694
  4. 2014Very Deep Convolutional Networks for Large-Scale Image Recognitioncited by 15,267
  5. 2021Diffusion Models Beat GANs on Image Synthesiscited by 8,185
  6. 2014Explaining and Harnessing Adversarial Examplescited by 7,287
  7. 2016Densely Connected Convolutional Networkscited by 6,691
  8. 2022Hierarchical Text-Conditional Image Generation with CLIP Latentscited by 6,539
  9. 2022Classifier-Free Diffusion Guidancecited by 5,717
  10. 2015Rethinking the Inception Architecture for Computer Visioncited by 5,451
  11. 2017mixup: Beyond Empirical Risk Minimizationcited by 5,155
  12. 2016Wide Residual Networkscited by 3,057
38 more
  1. 2022Elucidating the Design Space of Diffusion-Based Generative Modelscited by 2,994
  2. 2019CutMix: Regularization Strategy to Train Strong Classifiers with Localizable Featurescited by 2,875
  3. 2019Benchmarking Neural Network Robustness to Common Corruptions and Perturbationscited by 2,857
  4. 2016Perceptual Losses for Real-Time Style Transfer and Super-Resolutioncited by 2,678
  5. 2016Aggregated Residual Transformations for Deep Neural Networkscited by 2,540
  6. 2017Arbitrary Style Transfer in Real-time with Adaptive Instance Normalizationcited by 1,998
  7. 2020The Many Faces of Robustness: A Critical Analysis of Out-of-Distribution Generalizationcited by 1,842
  8. 2018ImageNet-trained CNNs are biased towards texture; increasing shape bias improves accuracy and robustnesscited by 1,430
  9. 2018Averaging Weights Leads to Wider Optima and Better Generalizationcited by 1,194
  10. 2016Instance Normalization: The Missing Ingredient for Fast Stylizationcited by 1,187
  11. 2019AugMix: A Simple Data Processing Method to Improve Robustness and Uncertaintycited by 897
  12. 2018Robustness May Be at Odds with Accuracycited by 873
  13. 2020RobustBench: a standardized adversarial robustness benchmarkcited by 659
  14. 2015A Neural Algorithm of Artistic Stylecited by 480
  15. 2017The Effectiveness of Data Augmentation in Image Classification using Deep Learningcited by 368
  16. 2017Population Based Training of Neural Networkscited by 351
  17. 2016A Learned Representation For Artistic Stylecited by 340
  18. 2023Synthetic Data from Diffusion Models Improves ImageNet Classificationcited by 326
  19. 2022Is synthetic data from generative models ready for image recognition?cited by 320
  20. 2017Unlabeled Samples Generated by GAN Improve the Person Re-identification Baseline in vitrocited by 280
  21. 2020Uncovering the Limits of Adversarial Training against Norm-Bounded Adversarial Examplescited by 275
  22. 2018GAN-based Synthetic Medical Image Augmentation for increased CNN Performance in Liver Lesion Classificationcited by 242
  23. 2021Improving Robustness using Generated Datacited by 233
  24. 2021TrivialAugment: Tuning-free Yet State-of-the-Art Data Augmentationcited by 225
  25. 2023Better Diffusion Models Further Improve Adversarial Trainingcited by 207
  26. 2021Fixing Data Augmentation to Improve Adversarial Robustnesscited by 201
  27. 2019Population Based Augmentation: Efficient Learning of Augmentation Policy Schedulescited by 153
  28. 2021PixMix: Dreamlike Pictures Comprehensively Improve Safety Measurescited by 131
  29. 2023DatasetDM: Synthesizing Data with Perception Annotations Using Diffusion Modelscited by 122
  30. 2019Improving Robustness Without Sacrificing Accuracy with Patch Gaussian Augmentationcited by 105
  31. 2021On Interaction Between Augmentations and Corruptions in Natural Corruption Robustnesscited by 98
  32. 2017Exploring the structure of a real-time, arbitrary neural artistic stylization networkcited by 89
  33. 2018Style Augmentation: Data Augmentation via Style Randomizationcited by 79
  34. 2021Noisy Feature Mixupcited by 27
  35. 2022NoisyMix: Boosting Model Robustness to Common Corruptionscited by 17
  36. 2019Statistically Robust Neural Network Classificationcited by 16
  37. 2023IPMix: Label-Preserving Data Augmentation Method for Training Robust Classifierscited by 14
  38. 2023Investigating the Corruption Robustness of Image Classifiers with Random Lp-norm Corruptionscited by 4

50 most cited of 52 All 52 (JSON)

Built on by

Not cited yet.