Overlap-guided Gaussian Mixture Models for Point Cloud Registration

TL;DR

This paper introduces an overlap-guided probabilistic method for 3D point cloud registration that uses Gaussian Mixture Models and a Transformer-based detection module to improve accuracy and efficiency, especially with partial overlaps.

Contribution

It presents a novel overlap-guided registration approach combining GMMs and a Transformer-based overlap detection module, addressing partial overlap challenges.

Findings

Achieves superior accuracy over state-of-the-art methods.

Handles partial overlaps and varying densities effectively.

Demonstrates efficiency on synthetic and real-world datasets.

Abstract

Probabilistic 3D point cloud registration methods have shown competitive performance in overcoming noise, outliers, and density variations. However, registering point cloud pairs in the case of partial overlap is still a challenge. This paper proposes a novel overlap-guided probabilistic registration approach that computes the optimal transformation from matched Gaussian Mixture Model (GMM) parameters. We reformulate the registration problem as the problem of aligning two Gaussian mixtures such that a statistical discrepancy measure between the two corresponding mixtures is minimized. We introduce a Transformer-based detection module to detect overlapping regions, and represent the input point clouds using GMMs by guiding their alignment through overlap scores computed by this detection module. Experiments show that our method achieves superior registration accuracy and efficiency than state-of-the-art methods when handling point clouds with partial overlap and different densities on synthetic and real-world datasets. https://github.com/gfmei/ogmm