精选论文集：三维点云有哪些必读的论文？

声明：转载本文请联系本文作者

三维点云是最重要的三维数据表达方式之一，SLAM、三维重建、机器人感知等多个领域，点云都是最简单、最普通的表达方式。三维点云处理在学习中需要参考大量的论文资料文献，但在众多的论文中，有哪些论文是必读的呢？

在『三维点云处理』课程上，黎嘉信老师为我们推荐了很多必读的论文，深蓝学院助教管郡智结合老师推荐的论文，为我们整理了24篇精选论文，并根据三维点云应用的不同方向分类，希能为大家提供一定的帮助。

——文末有论文领取通道哦~——

一起来看看都有哪些论文吧！

1. Classification and Segmentation

(1) PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation. CR Qi et.al.
(2) PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Space.CR Qi et.al.
(3) SO-Net: Self-Organizing Network for Point Cloud http://Analysis.Li J , Chen B M , Lee G H

2. Object Detection• Multi-view Projection

(1) Multi-View 3D Object Detection Network for Autonomous Driving, Chen Xiaozhi, et. al.

• Voxel-based
(1) VoxelNet: End-to-End Learning for Point Cloud Based 3D Object Detection, Yin Zhou, et.al.
(2) SECOND: Sparsely Embedded Convolutional Detection.Yan Y , Mao Y , Li B .
(3) PointPillars: Fast Encoders for Object Detection from Point Clouds. AH Lang, et.al.

• Point-based
(1) PointRCNN:3D Object Proposal Generation and Detection from Point Cloud, S Shi, et.al.
(2) From Points to Parts: 3D Object Detection from Point Cloud with Part-aware and Part-aggregation Network.Shi S, Wang Z, Shi J, et.al.

• Voxel-based & Point-based fusion
(1) PV-RCNN: Improving 3D Object Detection with Learned Deformations.Bhattacharyya P , Czarnecki K . Deformable

• Point cloud & image fusion
(1) Frustum PointNets for 3D Object Detection from RGB-D Data, Charles R. Qi, et.al.
(2) PointPainting: Sequential Fusion for 3D Object Detection, Sourabh Vora, et.al.

3. Point Cloud Registration

• Keypoints detection

A. Classic algorithm
(1) Harris 3D: a robust extension of the Harris operator for interest point detection on 3D meshes.Sipiran I , Bustos B .
(2) Intrinsic shape signatures: A shape descriptor for 3D object recognition[C]// IEEE International Conference on Computer Vision Workshops. Zhong Y.
B. Deep learning
(1) USIP: Unsupervised Stable Interest Point Detection from 3D Point http://Clouds.Li J , Lee G H .
(2) 3DFeat-Net: Weakly Supervised Local 3D Features for Point Cloud Registration[C]// European Conference on Computer Vision. Yew Z J , Lee G H .

• Discriptor

A. Classic algorithm
(1) Fast Point Feature Histograms (FPFH) for 3D registration[C]// IEEE International Conference on Robotics & Automation. Rusu R B , Blodow N , Beetz M .
(2) SHOT: Unique signatures of histograms for surface and texture description.Salti S , Tombari F , Stefano L D .
B. Deep learning
(1) 3DMatch: Learning Local Geometric Descriptors from RGB-D Reconstructions
(2) The Perfect Match: 3D Point Cloud Matching with Smoothed Densities
(3) PPFNet: Global Context Aware Local Features for Robust 3D Point Matching
(4) PPF-FoldNet: Unsupervised Learning of Rotation Invariant 3D Local Descriptors
(5) CGF: Learning Compact Geometric Features

• Regitrstion

(1) Least-Squares Fitting of Two 3-D Point Sets.Arun K S .
(2) The Normal Distributions Transform: A New Approach to Laser Scan Matching, Peter Biber, Wolfgang Straber, IROS 2003.

感谢深蓝学院助教管郡智同学的整理。希望这些精选的论文可以为大家的学习提供一定的帮助。

精选论文已备好，添加黎新“csig-igal”备注“点云论文集”，即可免费领取。