Domain-Agnostic Prior for Transfer Semantic Segmentation

TL;DR

This paper introduces a domain-agnostic prior (DAP) to improve unsupervised domain adaptation in semantic segmentation by aligning features with a shared, modality-agnostic space, enhancing transfer performance across domains.

Contribution

The paper proposes a simple, effective DAP regularization method that improves UDA performance, especially using text embedding models as proxies, outperforming existing approaches.

Findings

DAP improves segmentation accuracy in UDA tasks.

Text embedding models serve as effective proxies for domain alignment.

Better proxies from various data modalities enhance UDA results.

Abstract

Unsupervised domain adaptation (UDA) is an important topic in the computer vision community. The key difficulty lies in defining a common property between the source and target domains so that the source-domain features can align with the target-domain semantics. In this paper, we present a simple and effective mechanism that regularizes cross-domain representation learning with a domain-agnostic prior (DAP) that constrains the features extracted from source and target domains to align with a domain-agnostic space. In practice, this is easily implemented as an extra loss term that requires a little extra costs. In the standard evaluation protocol of transferring synthesized data to real data, we validate the effectiveness of different types of DAP, especially that borrowed from a text embedding model that shows favorable performance beyond the state-of-the-art UDA approaches in terms of segmentation accuracy. Our research reveals that UDA benefits much from better proxies, possibly from other data modalities.