Unsupervised Domain Adaptation for Point Cloud Semantic Segmentation via Graph Matching

TL;DR

This paper introduces a graph-based unsupervised domain adaptation method for point cloud semantic segmentation that aligns local features and preserves semantic discrimination, outperforming existing global alignment approaches.

Contribution

The paper proposes a novel local-level feature alignment framework using graph matching and category-guided contrastive loss for improved domain adaptation in point cloud segmentation.

Findings

Achieves state-of-the-art results on synthetic-to-real and real-to-real adaptation tasks.

Effectively preserves semantic discrimination during domain transfer.

Outperforms methods based on global feature alignment.

Abstract

Unsupervised domain adaptation for point cloud semantic segmentation has attracted great attention due to its effectiveness in learning with unlabeled data. Most of existing methods use global-level feature alignment to transfer the knowledge from the source domain to the target domain, which may cause the semantic ambiguity of the feature space. In this paper, we propose a graph-based framework to explore the local-level feature alignment between the two domains, which can reserve semantic discrimination during adaptation. Specifically, in order to extract local-level features, we first dynamically construct local feature graphs on both domains and build a memory bank with the graphs from the source domain. In particular, we use optimal transport to generate the graph matching pairs. Then, based on the assignment matrix, we can align the feature distributions between the two domains with the graph-based local feature loss. Furthermore, we consider the correlation between the features of different categories and formulate a category-guided contrastive loss to guide the segmentation model to learn discriminative features on the target domain. Extensive experiments on different synthetic-to-real and real-to-real domain adaptation scenarios demonstrate that our method can achieve state-of-the-art performance.