Deep Domain Generalization with Feature-norm Network

TL;DR

This paper introduces feature-norm networks that improve domain generalization in image classification, especially under category shift, by avoiding negative transfer and enhancing robustness without domain-specific adaptation.

Contribution

The paper proposes a novel feature-norm network (FNN) and a collaborative version (CFNN) that enhance domain generalization without requiring label space matching, addressing category shift issues.

Findings

FNN outperforms existing domain generalization methods.

CFNN further improves accuracy by increasing posterior entropy.

Significant gains demonstrated on image classification benchmarks.

Abstract

In this paper, we tackle the problem of training with multiple source domains with the aim to generalize to new domains at test time without an adaptation step. This is known as domain generalization (DG). Previous works on DG assume identical categories or label space across the source domains. In the case of category shift among the source domains, previous methods on DG are vulnerable to negative transfer due to the large mismatch among label spaces, decreasing the target classification accuracy. To tackle the aforementioned problem, we introduce an end-to-end feature-norm network (FNN) which is robust to negative transfer as it does not need to match the feature distribution among the source domains. We also introduce a collaborative feature-norm network (CFNN) to further improve the generalization capability of FNN. The CFNN matches the predictions of the next most likely categories for each training sample which increases each network's posterior entropy. We apply the proposed FNN and CFNN networks to the problem of DG for image classification tasks and demonstrate significant improvement over the state-of-the-art.