HR-APR: APR-agnostic Framework with Uncertainty Estimation and Hierarchical Refinement for Camera Relocalisation

Changkun Liu1, Shuai Chen3, Yukun Zhao1, Huajian Huang1, Victor Prisacariu3, Tristan Braud1,2
HKUST CSE1, HKUST ISD2, Active Vision Lab, University of Oxford3
ICRA 2024
Paper Code arXiv

An overview of this paper. There is a typo in 48 seconds, should be "predictions have higher accuracy for queries close to the training set."

Abstract

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.

BibTeX

@inproceedings{liu2024hrapr,
        title = {HR-APR: APR-agnostic Framework with Uncertainty Estimation and Hierarchical Refinement for Camera Relocalisation},
        author={Changkun Liu and Shuai Chen and Yukun Zhao and Huajian Huang and Victor Prisacariu and Tristan Braud},
        booktitle = {International Conference on Robotics and Automation (ICRA)},
        year = {2024},
        organization={IEEE}
        }