HR-APR: APR-agnostic framework with uncertainty estimation and hierarchical refinement for camera relocalisation

Conference item

Absolute Pose Regressors (APRs) directly estimate camera poses from monocular images, but their accuracy is unstable for different queries. Uncertainty-aware APRs provide uncertainty information on the estimated pose, alleviating the impact of these unreliable predictions. However, existing uncertainty modelling techniques are often coupled with a specific APR architecture, resulting in suboptimal performance compared to state-of-the-art (SOTA) APR methods. This work introduces a novel APR-agnostic framework, HR-APR, that formulates uncertainty estimation as cosine similarity estimation between the query and database features. It does not rely on or affect APR network architecture, which is flexible and computationally efficient. In addition, we take advantage of the uncertainty for pose refinement to enhance the performance of APR. The extensive experiments demonstrate the effectiveness of our framework, reducing 27.4% and 15.2% of computational overhead on the 7Scenes and Cambridge Landmarks datasets while maintaining the SOTA accuracy in single-image APRs.

Authors: Liu, C; Chen, S; Zhao, Y; Huang, H; Prisacariu, V

• Publication status: Published
• Peer review status: Peer reviewed
• Publisher: IEEE
• Host title: 2024 IEEE International Conference on Robotics and Automation (ICRA)
• Pages: 8544-8550
• Publication date: 2024-08-08
• Acceptance date: 2024-01-29
• Event title: 2024 IEEE International Conference on Robotics and Automation (ICRA 2024)
• Event location: Yokohama, Japan
• Event website: https://2024.ieee-icra.org/
• Event start date: 2024-05-13
• Event end date: 2024-05-17
• DOI: 10.1109/icra57147.2024.10610903
• ISSN: 1050-4729
• EISBN: 9798350384574
• ISBN: 9798350384581
• Language: English
• Keywords: training; uncertainty; accuracy; databases; pipelines; estimation; network architecture