CVMar 25, 2025

Scene-agnostic Pose Regression for Visual Localization

arXiv:2503.19543v17 citationsh-index: 10CVPR
Originality Highly original
AI Analysis

This addresses the challenge of adaptable visual localization for robotics and AR/VR applications, representing a novel paradigm rather than an incremental improvement.

The paper tackles the problem of visual localization by introducing Scene-agnostic Pose Regression (SPR), which achieves accurate pose regression without retraining or databases, outperforming existing methods like APR, RPR, and VO in unknown scenes on the 360SPR and 360Loc datasets.

Absolute Pose Regression (APR) predicts 6D camera poses but lacks the adaptability to unknown environments without retraining, while Relative Pose Regression (RPR) generalizes better yet requires a large image retrieval database. Visual Odometry (VO) generalizes well in unseen environments but suffers from accumulated error in open trajectories. To address this dilemma, we introduce a new task, Scene-agnostic Pose Regression (SPR), which can achieve accurate pose regression in a flexible way while eliminating the need for retraining or databases. To benchmark SPR, we created a large-scale dataset, 360SPR, with over 200K photorealistic panoramas, 3.6M pinhole images and camera poses in 270 scenes at three different sensor heights. Furthermore, a SPR-Mamba model is initially proposed to address SPR in a dual-branch manner. Extensive experiments and studies demonstrate the effectiveness of our SPR paradigm, dataset, and model. In the unknown scenes of both 360SPR and 360Loc datasets, our method consistently outperforms APR, RPR and VO. The dataset and code are available at https://junweizheng93.github.io/publications/SPR/SPR.html.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes