Hallucinating Agnostic Images to Generalize Across Domains

TL;DR

This paper introduces a method to generate agnostic images that retain category information while removing style specifics, enabling better domain generalization without requiring target labels.

Contribution

It proposes a novel deep architecture that hallucinate agnostic images using pixel adaptation and adversarial training, improving multi-source domain adaptation and generalization.

Findings

Effective in multi-source domain adaptation

Enhances domain generalization performance

Works with only source data and unlabeled target samples

Abstract

The ability to generalize across visual domains is crucial for the robustness of artificial recognition systems. Although many training sources may be available in real contexts, the access to even unlabeled target samples cannot be taken for granted, which makes standard unsupervised domain adaptation methods inapplicable in the wild. In this work we investigate how to exploit multiple sources by hallucinating a deep visual domain composed of images, possibly unrealistic, able to maintain categorical knowledge while discarding specific source styles. The produced agnostic images are the result of a deep architecture that applies pixel adaptation on the original source data guided by two adversarial domain classifier branches at image and feature level. Our approach is conceived to learn only from source data, but it seamlessly extends to the use of unlabeled target samples. Remarkable results for both multi-source domain adaptation and domain generalization support the power of hallucinating agnostic images in this framework.