MatchNorm: Learning-based Point Cloud Registration for 6D Object Pose Estimation in the Real World

TL;DR

This paper introduces MatchNorm, a normalization strategy and a new loss function to improve learning-based 3D point cloud registration for 6D object pose estimation in real-world data, addressing previous failures on real datasets.

Contribution

The paper proposes a novel normalization method and a loss function that enhance the robustness of learning-based registration methods on real-world point cloud data.

Findings

Improved registration accuracy on real-world datasets

Enabling learning-based methods to succeed in real-world scenarios

Generalizable strategies applicable to existing frameworks

Abstract

In this work, we tackle the task of estimating the 6D pose of an object from point cloud data. While recent learning-based approaches to addressing this task have shown great success on synthetic datasets, we have observed them to fail in the presence of real-world data. We thus analyze the causes of these failures, which we trace back to the difference between the feature distributions of the source and target point clouds, and the sensitivity of the widely-used SVD-based loss function to the range of rotation between the two point clouds. We address the first challenge by introducing a new normalization strategy, Match Normalization, and the second via the use of a loss function based on the negative log likelihood of point correspondences. Our two contributions are general and can be applied to many existing learning-based 3D object registration frameworks, which we illustrate by implementing them in two of them, DCP and IDAM. Our experiments on the real-scene TUD-L, LINEMOD and Occluded-LINEMOD datasets evidence the benefits of our strategies. They allow for the first time learning-based 3D object registration methods to achieve meaningful results on real-world data. We therefore expect them to be key to the future development of point cloud registration methods.