Improving Out-of-Domain Robustness with Targeted Augmentation in Frequency and Pixel Spaces

TL;DR

This paper introduces Frequency-Pixel Connect, a dataset-agnostic targeted augmentation method in frequency and pixel spaces that improves out-of-domain robustness across diverse real-world benchmarks.

Contribution

It proposes a novel, dataset-agnostic targeted augmentation framework that enhances OOD robustness by mixing amplitude spectrum and pixel content, outperforming existing methods.

Findings

Significantly improves cross-domain connectivity.

Outperforms previous generic and dataset-specific augmentation methods.

Effective across vision, medical, audio, and astronomical datasets.

Abstract

Out-of-domain (OOD) robustness under domain adaptation settings, where labeled source data and unlabeled target data come from different distributions, is a key challenge in real-world applications. A common approach to improving OOD robustness is through data augmentations. However, in real-world scenarios, models trained with generic augmentations can only improve marginally when generalized under distribution shifts toward unlabeled target domains. While dataset-specific targeted augmentations can address this issue, they typically require expert knowledge and extensive prior data analysis to identify the nature of the datasets and domain shift. To address these challenges, we propose Frequency-Pixel Connect, a domain-adaptation framework that enhances OOD robustness by introducing a targeted augmentation in both the frequency space and pixel space. Specifically, we mix the amplitude spectrum and pixel content of a source image and a target image to generate augmented samples that introduce domain diversity while preserving the semantic structure of the source image. Unlike previous targeted augmentation methods that are both dataset-specific and limited to the pixel space, Frequency-Pixel Connect is dataset-agnostic, enabling broader and more flexible applicability beyond natural image datasets. We further analyze the effectiveness of Frequency-Pixel Connect by evaluating the performance of our method connecting same-class cross-domain samples while separating different-class examples. We demonstrate that Frequency-Pixel Connect significantly improves cross-domain connectivity and outperforms previous generic methods on four diverse real-world benchmarks across vision, medical, audio, and astronomical domains, and it also outperforms other dataset-specific targeted augmentation methods.