MultiLoc: Multi-view Guided Relative Pose Regression for Fast and Robust Visual Re-Localization

2026-03-28

AI Summary

Existing relative pose regression methods generalize well to unseen environments but suffer from limited localization accuracy due to their reliance on pairwise, local viewpoints. This work proposes MultiLoc, a multi-view guided relative pose regression model that achieves efficient and accurate zero-shot pose estimation in a single forward pass by leveraging co-visibility-driven reference view retrieval and a multi-view joint feature fusion mechanism. By endowing the model with globally consistent spatial and geometric understanding, MultiLoc significantly enhances cross-domain robustness. The method outperforms current state-of-the-art approaches on benchmarks including WaySpots, Cambridge Landmarks, and Indoor6, and achieves leading performance in relative pose estimation on MegaDepth-1500, ScanNet-1500, and ACID.

Abstract

Relative Pose Regression (RPR) generalizes well to unseen environments, but its performance is often limited due to pairwise and local spatial views. To this end, we propose MultiLoc, a novel multi-view guided RPR model trained at scale, equipping relative pose regression with globally consistent spatial and geometric understanding. Specifically, our method jointly fuses multiple reference views and their associated camera poses in a single forward pass, enabling accurate zero-shot pose estimation with real-time efficiency. To reliably supply informative context, we further propose a co-visibility-driven retrieval strategy for geometrically relevant reference view selection. MultiLoc establishes a new benchmark in visual re-localization, consistently outperforming existing state-of-the-art (SOTA) relative pose regression (RPR) methods across diverse datasets, including WaySpots, Cambridge Landmarks, and Indoor6. Furthermore, MultiLoc's pose regressor exhibits SOTA performance in relative pose estimation, surpassing RPR, feature matching and non-regression-based techniques on the MegaDepth-1500, ScanNet-1500, and ACID benchmarks. These results demonstrate robust domain generalization of MultiLoc across indoor, outdoor and natural environments. Code will be made publicly available.

Problem

• Relative Pose Regression
• Visual Re-localization
• Multi-view
• Pose Estimation
• Domain Generalization

Innovation

• multi-view fusion
• relative pose regression
• co-visibility retrieval
• zero-shot localization
• visual re-localization