Robust Camera Pose Refinement for Multi-Resolution Hash Encoding

TL;DR

This paper introduces a joint optimization method for camera pose refinement and geometric representation learning using multi-resolution hash encoding, improving neural rendering accuracy and convergence speed.

Contribution

It proposes a novel joint optimization algorithm that stabilizes gradient oscillations and employs curriculum training, enhancing pose refinement with hash encoding.

Findings

Achieves state-of-the-art performance on novel-view synthesis datasets.

Enables rapid convergence even with unknown initial camera poses.

Effectively stabilizes gradient flows in hash encoding for pose calibration.

Abstract

Multi-resolution hash encoding has recently been proposed to reduce the computational cost of neural renderings, such as NeRF. This method requires accurate camera poses for the neural renderings of given scenes. However, contrary to previous methods jointly optimizing camera poses and 3D scenes, the naive gradient-based camera pose refinement method using multi-resolution hash encoding severely deteriorates performance. We propose a joint optimization algorithm to calibrate the camera pose and learn a geometric representation using efficient multi-resolution hash encoding. Showing that the oscillating gradient flows of hash encoding interfere with the registration of camera poses, our method addresses the issue by utilizing smooth interpolation weighting to stabilize the gradient oscillation for the ray samplings across hash grids. Moreover, the curriculum training procedure helps to learn the level-wise hash encoding, further increasing the pose refinement. Experiments on the novel-view synthesis datasets validate that our learning frameworks achieve state-of-the-art performance and rapid convergence of neural rendering, even when initial camera poses are unknown.