# Downscaling Attack and Defense: Turning What You See Back Into What You Get

arXiv: https://arxiv.org/abs/2010.02456 · HTML: https://mlsanity.com/paper/2010.02456/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 6 Oct 2020.

## Built on by (most cited first, 1 of 1)

- [FrequencyLowCut Pooling -- Plug & Play against Catastrophic Overfitting](https://mlsanity.com/paper/2204.00491/) (2022)

## Abstract (arXiv)

The resizing of images, which is typically a required part of preprocessing for computer vision systems, is vulnerable to attack. Images can be created such that the image is completely different at machine-vision scales than at other scales and the default settings for some common computer vision and machine learning systems are vulnerable. We show that defenses exist and are trivial to administer provided that defenders are aware of the threat. These attacks and defenses help to establish the role of input sanitization in machine learning.