A Novel Cross-Perturbation for Single Domain Generalization

TL;DR

This paper introduces CPerb, a cross-perturbation method combining image and feature-level augmentations with multi-route perturbation and MixPatch to improve single domain generalization by increasing data diversity and learning domain-invariant features.

Contribution

The paper proposes CPerb, a novel cross-perturbation approach that synergistically combines image and feature-level perturbations with multi-route strategies and introduces MixPatch for enhanced data diversity.

Findings

CPerb significantly improves generalization performance on benchmark datasets.

The combination of image and feature perturbations outperforms existing methods.

MixPatch further enhances model robustness and diversity.

Abstract

Single domain generalization aims to enhance the ability of the model to generalize to unknown domains when trained on a single source domain. However, the limited diversity in the training data hampers the learning of domain-invariant features, resulting in compromised generalization performance. To address this, data perturbation (augmentation) has emerged as a crucial method to increase data diversity. Nevertheless, existing perturbation methods often focus on either image-level or feature-level perturbations independently, neglecting their synergistic effects. To overcome these limitations, we propose CPerb, a simple yet effective cross-perturbation method. Specifically, CPerb utilizes both horizontal and vertical operations. Horizontally, it applies image-level and feature-level perturbations to enhance the diversity of the training data, mitigating the issue of limited diversity in single-source domains. Vertically, it introduces multi-route perturbation to learn domain-invariant features from different perspectives of samples with the same semantic category, thereby enhancing the generalization capability of the model. Additionally, we propose MixPatch, a novel feature-level perturbation method that exploits local image style information to further diversify the training data. Extensive experiments on various benchmark datasets validate the effectiveness of our method.