MoireMix: A Formula-Based Data Augmentation for Improving Image Classification Robustness

TL;DR

MoireMix introduces a fast, formula-based data augmentation technique using Moire interference patterns to enhance image classification robustness without external data or high computational costs.

Contribution

It presents a novel, lightweight, and storage-free augmentation method based on analytic interference patterns, outperforming existing approaches in robustness benchmarks.

Findings

Significantly improves robustness on ImageNet-C and ImageNet-R

Achieves negligible computational overhead (0.0026 seconds per image)

Outperforms standard and external-data-free augmentation methods

Abstract

Data augmentation is a key technique for improving the robustness of image classification models. However, many recent approaches rely on diffusion-based synthesis or complex feature mixing strategies, which introduce substantial computational overhead or require external datasets. In this work, we explore a different direction: procedural augmentation based on analytic interference patterns. Unlike conventional augmentation methods that rely on stochastic noise, feature mixing, or generative models, our approach exploits Moire interference to generate structured perturbations spanning a wide range of spatial frequencies. We propose a lightweight augmentation method that procedurally generates Moire textures on-the-fly using a closed-form mathematical formulation. The patterns are synthesized directly in memory with negligible computational cost (0.0026 seconds per image), mixed with training images during training, and immediately discarded, enabling a storage-free augmentation pipeline without external data. Extensive experiments with Vision Transformers demonstrate that the proposed method consistently improves robustness across multiple benchmarks, including ImageNet-C, ImageNet-R, and adversarial benchmarks, outperforming standard augmentation baselines and existing external-data-free augmentation approaches. These results suggest that analytic interference patterns provide a practical and efficient alternative to data-driven generative augmentation methods.