ReLoc: Rethinking Scene Coordinate Regression Architecture for Robust Outdoor LiDAR-based Localization

2026-09-27

AI Summary

This study addresses key limitations of existing scene coordinate regression methods in LiDAR-based localization, specifically the difficulty in distinguishing proximate locations and the dynamic noise introduced by uniform sampling that disrupts training. To overcome these challenges, this work proposes ReLoc, a novel architecture that optimizes global embeddings via learnable context tokens and enhances local features through an attention-driven feature aggregator. This design effectively suppresses noise while reinforcing structural consistency, thereby enabling fine-grained spatial discrimination. Extensive evaluations on two large-scale outdoor datasets demonstrate that the proposed method achieves state-of-the-art localization accuracy while maintaining real-time inference performance.

Abstract

Scene Coordinate Regression (SCR) has recently emerged as a promising approach for LiDAR-based localization, achieving accurate localization without requiring an explicit 3D map. Despite their effectiveness, existing SCR methods rely on scene classification-based global embedding that struggles to provide fine-grained discrimination among nearby locations. Moreover, their reliance on uniform sampling of local features during training assigns equal importance to all points, thereby inadvertently propagating features from dynamic objects or unstable regions and potentially degrading training stability. In this paper, we present ReLoc, a revamped SCR architecture that can effectively address these limitations. First, we redesign the global embedding module by combining learnable context tokens with a feature aggregator to capture richer and more discriminative scene context. Second, we introduce an attention-based local feature enhancement module to mitigate the impact of noisy local features while encouraging context-consistent structures, yielding more robust local feature representations. Experimental results on two large-scale outdoor datasets demonstrate that our approach achieves state-of-the-art accuracy over previous SCR-based methods while maintaining real-time inference performance.

Problem

• Scene Coordinate Regression
• LiDAR-based Localization
• Global Embedding
• Local Feature Sampling
• Outdoor Localization

Innovation

• Scene Coordinate Regression
• LiDAR Localization
• Global Embedding
• Attention Mechanism
• Local Feature Enhancement