Cross-view Domain Generalization via Geometric Consistency for LiDAR Semantic Segmentation

TL;DR

This paper introduces CVGC, a novel framework that enhances LiDAR semantic segmentation across different viewpoints by modeling geometric variations and enforcing consistency, significantly improving cross-view domain generalization.

Contribution

The paper proposes a cross-view geometric augmentation and consistency framework for LiDAR segmentation, addressing viewpoint differences and improving generalization to unseen domains.

Findings

CVGC outperforms state-of-the-art methods in cross-view scenarios.

Extensive experiments on six datasets validate the effectiveness of CVGC.

The approach is the first systematic evaluation of cross-view domain generalization for LiDAR segmentation.

Abstract

Domain-generalized LiDAR semantic segmentation (LSS) seeks to train models on source-domain point clouds that generalize reliably to multiple unseen target domains, which is essential for real-world LiDAR applications. However, existing approaches assume similar acquisition views (e.g., vehicle-mounted) and struggle in cross-view scenarios, where observations differ substantially due to viewpoint-dependent structural incompleteness and non-uniform point density. Accordingly, we formulate cross-view domain generalization for LiDAR semantic segmentation and propose a novel framework, termed CVGC (Cross-View Geometric Consistency). Specifically, we introduce a cross-view geometric augmentation module that models viewpoint-induced variations in visibility and sampling density, generating multiple cross-view observations of the same scene. Subsequently, a geometric consistency module enforces consistent semantic and occupancy predictions across geometrically augmented point clouds of the same scene. Extensive experiments on six public LiDAR datasets establish the first systematic evaluation of cross-view domain generalization for LiDAR semantic segmentation, demonstrating that CVGC consistently outperforms state-of-the-art methods when generalizing from a single source domain to multiple target domains with heterogeneous acquisition viewpoints. The source code will be publicly available at https://github.com/KintomZi/CVGC-DG