# An Effective Anti-Aliasing Approach for Residual Networks

arXiv: https://arxiv.org/abs/2011.10675 · HTML: https://mlsanity.com/paper/2011.10675/

## Key facts

- Cited by 23 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 20 Nov 2020.

## Built on by (most cited first, 23 of 23)

- [Alias-Free Generative Adversarial Networks](https://mlsanity.com/paper/2106.12423/) (2021)
- [Towards Robust Vision Transformer](https://mlsanity.com/paper/2105.07926/) (2021)
- [AugMax: Adversarial Composition of Random Augmentations for Robust Training](https://mlsanity.com/paper/2110.13771/) (2021)
- [MedViT: A Robust Vision Transformer for Generalized Medical Image Classification](https://mlsanity.com/paper/2302.09462/) (2023)
- [Deep networks for system identification: a Survey](https://mlsanity.com/paper/2301.12832/) (2023)
- [Taxonomy of Machine Learning Safety: A Survey and Primer](https://mlsanity.com/paper/2106.04823/) (2021)
- [Blending Anti-Aliasing into Vision Transformer](https://mlsanity.com/paper/2110.15156/) (2021)
- [Blurs Behave Like Ensembles: Spatial Smoothings to Improve Accuracy, Uncertainty, and Robustness](https://mlsanity.com/paper/2105.12639/) (2021)
- [How Does Frequency Bias Affect the Robustness of Neural Image Classifiers against Common Corruption and Adversarial Perturbations?](https://mlsanity.com/paper/2205.04533/) (2022)
- [Alias-Free Convnets: Fractional Shift Invariance via Polynomial Activations](https://mlsanity.com/paper/2303.08085/) (2023)
- [How Convolutional Neural Networks Deal with Aliasing](https://mlsanity.com/paper/2102.07757/) (2021)
- [Using Synthetic Corruptions to Measure Robustness to Natural Distribution Shifts](https://mlsanity.com/paper/2107.12052/) (2021)
- [Bridging the Gap Between Adversarial Robustness and Optimization Bias](https://mlsanity.com/paper/2102.08868/) (2021)
- [Improving Sound Event Classification by Increasing Shift Invariance in Convolutional Neural Networks](https://mlsanity.com/paper/2107.00623/) (2021)
- [Improving Model Generalization by Agreement of Learned Representations from Data Augmentation](https://mlsanity.com/paper/2110.10536/) (2021)
- [Combining Different V1 Brain Model Variants to Improve Robustness to Image Corruptions in CNNs](https://mlsanity.com/paper/2110.10645/) (2021)
- [Enhancing object detection robustness: A synthetic and natural perturbation approach](https://mlsanity.com/paper/2304.10622/) (2023)
- [On the Shift Invariance of Max Pooling Feature Maps in Convolutional Neural Networks](https://mlsanity.com/paper/2209.11740/) (2022)
- [The Importance of Anti-Aliasing in Tiny Object Detection](https://mlsanity.com/paper/2310.14221/) (2023)
- [PushPull-Net: Inhibition-driven ResNet robust to image corruptions](https://mlsanity.com/paper/2408.04077/) (2024)
- [Alias-Free ViT: Fractional Shift Invariance via Linear Attention](https://mlsanity.com/paper/2510.22673/) (2025)
- [MoireDB: Formula-generated Interference-fringe Image Dataset](https://mlsanity.com/paper/2502.01490/) (2025)
- [WaveInst: A Frequency-Domain Enhanced Network for Fine-Grained Thin Tree Trunk Extraction in Forest Scenes](https://mlsanity.com/paper/2505.01656/) (2025)

## Abstract (arXiv)

Image pre-processing in the frequency domain has traditionally played a vital role in computer vision and was even part of the standard pipeline in the early days of deep learning. However, with the advent of large datasets, many practitioners concluded that this was unnecessary due to the belief that these priors can be learned from the data itself. Frequency aliasing is a phenomenon that may occur when sub-sampling any signal, such as an image or feature map, causing distortion in the sub-sampled output. We show that we can mitigate this effect by placing non-trainable blur filters and using smooth activation functions at key locations, particularly where networks lack the capacity to learn them. These simple architectural changes lead to substantial improvements in out-of-distribution generalization on both image classification under natural corruptions on ImageNet-C [10] and few-shot learning on Meta-Dataset [17], without introducing additional trainable parameters and using the default hyper-parameters of open source codebases.