DSDRNet: Disentangling Representation and Reconstruct Network for Domain Generalization

TL;DR

This paper introduces DSDRNet, a novel disentanglement-reconstruction network that improves domain generalization by combining inter- and intra-instance features with adaptive normalization, outperforming existing methods on benchmark datasets.

Contribution

The paper proposes a dual-stream disentanglement and reconstruction approach with novel supervised signals and adaptive normalization, advancing domain generalization techniques without relying on domain labels.

Findings

DSDRNet outperforms existing methods on four benchmark datasets.

The dual-stream fusion effectively captures inter- and intra-instance features.

Adaptive Instance Normalization enhances model convergence and reconstruction quality.

Abstract

Domain generalization faces challenges due to the distribution shift between training and testing sets, and the presence of unseen target domains. Common solutions include domain alignment, meta-learning, data augmentation, or ensemble learning, all of which rely on domain labels or domain adversarial techniques. In this paper, we propose a Dual-Stream Separation and Reconstruction Network, dubbed DSDRNet. It is a disentanglement-reconstruction approach that integrates features of both inter-instance and intra-instance through dual-stream fusion. The method introduces novel supervised signals by combining inter-instance semantic distance and intra-instance similarity. Incorporating Adaptive Instance Normalization (AdaIN) into a two-stage cyclic reconstruction process enhances self-disentangled reconstruction signals to facilitate model convergence. Extensive experiments on four benchmark datasets demonstrate that DSDRNet outperforms other popular methods in terms of domain generalization capabilities.