PointCloud-Registration-Tracing

Inspired by awesome-point-cloud-registration and awesome-computer-vision.

This list focuses on the rigid registration between point clouds and tracks progress in point cloud registration research.

Table of Contents

• 1.Overview
• 2.Coarse Registration (Global Registration)
  • 2.1 Feature Matching Based
    • 2.1.1 Feature Description
    • 2.1.2 Keypoint Detection
    • 2.1.3 Outlier Rejection
    • 2.1.4 Graph Algorithms
  • 2.2 End-to-End
  • 2.3 Randomized
  • 2.4 Probablistic
• 3.Fine Registration (Local Registration)
  • 3.1 Learning-based
  • 3.2 Traditional
• 4.Datasets
• 5.Tools

1.Overview

• Deep learning based point cloud registration: an overview 【VRIH, 2020】
• A comprehensive survey on point cloud registration 【arXiv, 2021】
• Transformers in 3D Point Clouds: A Survey 【arXiv, 2022】
• A review of non-rigid transformations and learning-based 3D point cloud registration methods【ISPRS, 2023】

2.Coarse Registration

2.1 Feature Matching Based

2.1.1 Feature Description

(1) Point-based Strategy

• PPFNet: Global Context Aware Local Features for Robust 3D Point Matching【CVPR, 2018】
• PPF-FoldNet: Unsupervised Learning of Rotation Invariant 3D Local Descriptors【ECCV, 2018】【code】
• 3DFeat-Net: Weakly Supervised Local 3D Features for Point Cloud Registration【ECCV, 2018】【code】
• D3Feat: Joint Learning of Dense Detection and Description of 3D Local Features 【CVPR, 2020】【code】

(2) Voxel-based Strategy

• 3DMatch: Learning Local Geometric Descriptors from RGB-D Reconstructions【CVPR, 2017】 【code】
• The Perfect Match: 3D Point Cloud Matching with Smoothed Densities【CVPR, 2019] 【code】
• FCGF: Fully Convolutional Geometric Features 【ICCV, 2019】【code】

(3) Point-Voxel Strategy

(4) Line-based Strategy

• SuperLine3D: Self-supervised Line Segmentation and Description for LiDAR Point Cloud【ECCV, 2022】【code】

(5) Transformer-based Strategy

• Full Transformer Framework for Robust Point Cloud Registration with Deep Information Interaction【arXiv, 2021】
• Geometric Transformer for Fast and Robust Point Cloud Registration【CVPR, 2022】【code】

(6) Other Strategy

• SpinNet: Learning a General Surface Descriptor for 3D Point Cloud Registration【CVPR, 2021】【code】
• You Only Hypothesize Once: Point Cloud Registration with Rotation-equivariant Descriptors【ICCV, 2021】【code】

2.1.2 Keypoint Detection

• USIP: Unsupervised Stable Interest Point Detection From 3D Point Clouds【ICCV, 2019】【code】
• Point cloud saliency detection by local and global feature fusion【T-IP, 2019】
• D3Feat: Joint Learning of Dense Detection and Description of 3D Local Features 【CVPR, 2020】【code】
• 3D3L: Deep learned 3D keypoint detection and description for lidars【ICRA, 2021】【code】
• UKPGAN: A General Self-Supervised Keypoint Detector【CVPR, 2022】【code】

2.1.3 Outlier Rejection

• RANSAC: Random Sample Consensus: A Paradigm for Model Fitting with Applications to Image Analysis and Automated Cartography 【1981】
• TEASER: Fast and Certifiable Point Cloud Registration 【T-RO, 2020】【code】
• SC^2-PCR: A Second Order Spatial Compatibility for Efficient and Robust Point Cloud Registration【CVPR, 2022】【code】

3.Fine Registration (Local Registration)

4.Datasets

• 3DMatch
• KITTI Odometry
• ETH (Challenging data sets for point cloud registration algorithms)
• WHU-TLS Benchmark
• ModelNet
• A Benchmark for Point Clouds Registration Algorithms
• Standford 3DScanning

5.Tools

• Open3D: A Modern Libray for 3D Data Processing
• PCL: Point Cloud Library
• Pytorch Geometric
• PyTorch Points 3D
• probreg