# OA-CutMix: Correcting the Label Bias of CutMix

arXiv: https://arxiv.org/abs/2606.04820 · HTML: https://mlsanity.com/paper/2606.04820/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 24 of its 33 references that are in the map.
- First posted on arXiv 3 Jun 2026.

## Builds on (most cited first, 24 of 24)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [mixup: Beyond Empirical Risk Minimization](https://mlsanity.com/paper/1710.09412/) (2017)
- [On Calibration of Modern Neural Networks](https://mlsanity.com/paper/1706.04599/) (2017)
- [Training data-efficient image transformers & distillation through attention](https://mlsanity.com/paper/2012.12877/) (2020)
- [CutMix: Regularization Strategy to Train Strong Classifiers with Localizable Features](https://mlsanity.com/paper/1905.04899/) (2019)
- [RandAugment: Practical automated data augmentation with a reduced search space](https://mlsanity.com/paper/1909.13719/) (2019)
- [Fine-Grained Visual Classification of Aircraft](https://mlsanity.com/paper/1306.5151/) (2013)
- [SAM 3: Segment Anything with Concepts](https://mlsanity.com/paper/2511.16719/) (2025)
- [DeiT III: Revenge of the ViT](https://mlsanity.com/paper/2204.07118/) (2022)
- [Effective Data Augmentation With Diffusion Models](https://mlsanity.com/paper/2302.07944/) (2023)
- [Puzzle Mix: Exploiting Saliency and Local Statistics for Optimal Mixup](https://mlsanity.com/paper/2009.06962/) (2020)
- [SaliencyMix: A Saliency Guided Data Augmentation Strategy for Better Regularization](https://mlsanity.com/paper/2006.01791/) (2020)
- [A Comprehensive Survey of Image Augmentation Techniques for Deep Learning](https://mlsanity.com/paper/2205.01491/) (2022)
- [ResizeMix: Mixing Data with Preserved Object Information and True Labels](https://mlsanity.com/paper/2012.11101/) (2020)
- [AutoMix: Unveiling the Power of Mixup for Stronger Classifiers](https://mlsanity.com/paper/2103.13027/) (2021)
- [SnapMix: Semantically Proportional Mixing for Augmenting Fine-grained Data](https://mlsanity.com/paper/2012.04846/) (2020)
- [TokenMix: Rethinking Image Mixing for Data Augmentation in Vision Transformers](https://mlsanity.com/paper/2207.08409/) (2022)
- [AlignMixup: Improving Representations By Interpolating Aligned Features](https://mlsanity.com/paper/2103.15375/) (2021)
- [Boosting Discriminative Visual Representation Learning with Scenario-Agnostic Mixup](https://mlsanity.com/paper/2111.15454/) (2021)
- [OpenMixup: Open Mixup Toolbox and Benchmark for Visual Representation Learning](https://mlsanity.com/paper/2209.04851/) (2022)
- [Harnessing Hard Mixed Samples with Decoupled Regularizer](https://mlsanity.com/paper/2203.10761/) (2022)
- [Data Augmentation for Image Classification using Generative AI](https://mlsanity.com/paper/2409.00547/) (2024)
- [ForAug: Mitigating Biases in Image Classification via Controlled Image Compositions](https://mlsanity.com/paper/2503.09399/) (2025)
- [Cut to the Mix: Simple Data Augmentation Outperforms Elaborate Ones in Limited Organ Segmentation Datasets](https://mlsanity.com/paper/2602.03555/) (2026)

## Abstract (arXiv)

CutMix has become the de facto standard mixing augmentation, yet its label assignment rests on a flawed assumption: The area of the pasted patch faithfully reflects its semantic contribution to the mixed image. In practice, however, patches frequently land on background regions, assigning label credit to classes whose objects are not visible. The mean discrepancy of the CutMix label and the semantic object area is 21.5%. In 17% of samples an image contributes zero visible object pixels yet receives nonzero label weight. We propose Object-Aware CutMix (OA-CutMix), which corrects this bias by replacing the area-based CutMix weight with one derived from precomputed segmentation masks, assigning labels in proportion to the visible object area each image contributes to the mix. The image mixing procedure is left entirely unchanged. We evaluate OA-CutMix against 10+ static and dynamic mixing methods across 4 architectures and 6 datasets. OA-CutMix consistently achieves the highest accuracy over all tasks, outperforming even dynamic mixing methods, but at a fraction of the training-time cost. Improvements are largest for small objects, where the label bias from CutMix is greatest. Thus, correcting the label is sufficient to match or exceed the performance of methods modifying the image mixing algorithm.