# Learning Visual Representations for Transfer Learning by Suppressing Texture

arXiv: https://arxiv.org/abs/2011.01901 · HTML: https://mlsanity.com/paper/2011.01901/

## Key facts

- Cited by 11 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 3 Nov 2020.

## Built on by (most cited first, 11 of 11)

- [Hyperbolic Contrastive Learning for Visual Representations beyond Objects](https://mlsanity.com/paper/2212.00653/) (2022)
- [Robust Contrastive Learning Using Negative Samples with Diminished Semantics](https://mlsanity.com/paper/2110.14189/) (2021)
- [Discrete Representations Strengthen Vision Transformer Robustness](https://mlsanity.com/paper/2111.10493/) (2021)
- [Object-Aware Cropping for Self-Supervised Learning](https://mlsanity.com/paper/2112.00319/) (2021)
- [On-target Adaptation](https://mlsanity.com/paper/2109.01087/) (2021)
- [ImageNet-trained CNNs are not biased towards texture: Revisiting feature reliance through controlled suppression](https://mlsanity.com/paper/2509.20234/) (2025)
- [Reducing Texture Bias of Deep Neural Networks via Edge Enhancing Diffusion](https://mlsanity.com/paper/2402.09530/) (2024)
- [Learning the Unlearned: Mitigating Feature Suppression in Contrastive Learning](https://mlsanity.com/paper/2402.11816/) (2024)
- [Shape Bias and Robustness Evaluation via Cue Decomposition for Image Classification and Segmentation](https://mlsanity.com/paper/2503.12453/) (2025)
- [Can We Break Free from Strong Data Augmentations in Self-Supervised Learning?](https://mlsanity.com/paper/2404.09752/) (2024)
- [Let Samples Speak: Mitigating Spurious Correlation by Exploiting the Clusterness of Samples](https://mlsanity.com/paper/2512.22874/) (2025)

## Abstract (arXiv)

Recent literature has shown that features obtained from supervised training of CNNs may over-emphasize texture rather than encoding high-level information. In self-supervised learning in particular, texture as a low-level cue may provide shortcuts that prevent the network from learning higher level representations. To address these problems we propose to use classic methods based on anisotropic diffusion to augment training using images with suppressed texture. This simple method helps retain important edge information and suppress texture at the same time. We empirically show that our method achieves state-of-the-art results on object detection and image classification with eight diverse datasets in either supervised or self-supervised learning tasks such as MoCoV2 and Jigsaw. Our method is particularly effective for transfer learning tasks and we observed improved performance on five standard transfer learning datasets. The large improvements (up to 11.49%) on the Sketch-ImageNet dataset, DTD dataset and additional visual analyses with saliency maps suggest that our approach helps in learning better representations that better transfer.