3D Point Cloud Registration with Multi-Scale Architecture and Unsupervised Transfer Learning

TL;DR

This paper introduces a novel 3D point cloud registration method combining multi-scale sparse voxel convolution and unsupervised transfer learning, achieving state-of-the-art results on real-world datasets.

Contribution

It presents a new deep learning framework with MS-SVConv and UDGE components for effective, unsupervised transfer learning in 3D point cloud registration tasks.

Findings

Achieves state-of-the-art registration accuracy on 3DMatch, ETH, and TUM datasets.

Demonstrates effective unsupervised transfer learning across diverse datasets.

Provides publicly available code for reproducibility.

Abstract

We propose a method for generalizing deep learning for 3D point cloud registration on new, totally different datasets. It is based on two components, MS-SVConv and UDGE. Using Multi-Scale Sparse Voxel Convolution, MS-SVConv is a fast deep neural network that outputs the descriptors from point clouds for 3D registration between two scenes. UDGE is an algorithm for transferring deep networks on unknown datasets in a unsupervised way. The interest of the proposed method appears while using the two components, MS-SVConv and UDGE, together as a whole, which leads to state-of-the-art results on real world registration datasets such as 3DMatch, ETH and TUM. The code is publicly available at https://github.com/humanpose1/MS-SVConv .