Category-Level Global Camera Pose Estimation with Multi-Hypothesis Point Cloud Correspondences

TL;DR

This paper introduces a novel registration method that maintains multiple possible correspondences and a new feature descriptor, significantly improving accuracy in category-level camera pose estimation, especially with noisy real-world data.

Contribution

It presents a new optimization approach that retains all potential correspondences and a local similarity-based feature descriptor, advancing category-level point cloud registration.

Findings

Outperforms state-of-the-art methods on various datasets.

Achieves up to 20% better performance on real-world noisy depth images.

Effective in matching objects within the same category.

Abstract

Correspondence search is an essential step in rigid point cloud registration algorithms. Most methods maintain a single correspondence at each step and gradually remove wrong correspondances. However, building one-to-one correspondence with hard assignments is extremely difficult, especially when matching two point clouds with many locally similar features. This paper proposes an optimization method that retains all possible correspondences for each keypoint when matching a partial point cloud to a complete point cloud. These uncertain correspondences are then gradually updated with the estimated rigid transformation by considering the matching cost. Moreover, we propose a new point feature descriptor that measures the similarity between local point cloud regions. Extensive experiments show that our method outperforms the state-of-the-art (SoTA) methods even when matching different objects within the same category. Notably, our method outperforms the SoTA methods when registering real-world noisy depth images to a template shape by up to 20% performance.