ML-SemReg: Boosting Point Cloud Registration with Multi-level Semantic Consistency

TL;DR

ML-SemReg is a novel point cloud registration framework that leverages multi-level semantic consistency to effectively handle low-overlap scenarios, significantly improving registration accuracy and robustness.

Contribution

The paper introduces a plug-and-play framework that exploits semantic information through group and mask matching modules to address inter- and intra-class mismatches in point cloud registration.

Findings

Achieves high inlier ratio in correspondences.

Improves registration recall by 34 percentage points on KITTI.

Demonstrates robustness in challenging cases.

Abstract

Recent advances in point cloud registration mostly leverage geometric information. Although these methods have yielded promising results, they still struggle with problems of low overlap, thus limiting their practical usage. In this paper, we propose ML-SemReg, a plug-and-play point cloud registration framework that fully exploits semantic information. Our key insight is that mismatches can be categorized into two types, i.e., inter- and intra-class, after rendering semantic clues, and can be well addressed by utilizing multi-level semantic consistency. We first propose a Group Matching module to address inter-class mismatching, outputting multiple matching groups that inherently satisfy Local Semantic Consistency. For each group, a Mask Matching module based on Scene Semantic Consistency is then introduced to suppress intra-class mismatching. Benefit from those two modules, ML-SemReg generates correspondences with a high inlier ratio. Extensive experiments demonstrate excellent performance and robustness of ML-SemReg, e.g., in hard-cases of the KITTI dataset, the Registration Recall of MAC increases by almost 34 percentage points when our ML-SemReg is equipped. Code is available at \url{https://github.com/Laka-3DV/ML-SemReg}