This repo collects papers on point cloud deep learning. Note that the stars I give to each paper contain personal bias for my own project, but actually I do appreciate all the works that have been done in this area. For my own purpose, I can't include all the papers that have been published. A more complete paper list since 2017 is here: https://github.com/Yochengliu/awesome-point-cloud-analysis.
- Escape from Cells: Deep Kd-Networks for the Recognition of 3D Point Cloud Models (ICCV 2017), R. Klokov et al. [pdf] ⭐ ⭐ ⭐ ⭐
- PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation (CVPR 2017), C. R. Qi et al. [pdf] [Github] ⭐ ⭐ ⭐ ⭐ ⭐
- PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Space (NeurIPS 2017), C. R. Qi et al. [pdf] [Github] ⭐ ⭐ ⭐ ⭐ ⭐
- PointCNN: Convolution On X-Transformed Points (NeurIPS 2018) Y. Li et al, [pdf] [Github] ⭐ ⭐ ⭐
- A-CNN: Annularly Convolutional Neural Networks on Point Clouds (CVPR 2019), A. Komarichev et al. [pdf]
⭐ ⭐ ⭐ - Relation-Shape Convolutional Neural Network for Point Cloud Analysis (CVPR 2019), Y. Liu et al. [pdf]
⭐ ⭐ ⭐ ⭐
Grid-based methods
- Voting for Voting in Online Point Cloud Object Detection (RSS 2015), D. Z. Wang et al. [pdf] ⭐ ⭐ ⭐
- Vote3Deep: Fast Object Detection in 3D Point Clouds Using Efficient Convolutional Neural Networks (ICRA 2017), M. Engelcke et al. [pdf] ⭐ ⭐ ⭐
- 3D fully convolutional network for vehicle detection in point cloud (IROS 2017) B. Li. [pdf] [Github] ⭐ ⭐ ⭐
- VoxelNet: End-to-End Learning for Point Cloud Based 3D Object Detection (CVPR 2018), Y. Zhou et al. [pdf]
⭐ ⭐ ⭐ ⭐ ⭐ - PIXOR: Real-time 3D Object Detection From Point Clouds (CVPR 2018), B. Yang et al. [pdf] ⭐ ⭐ ⭐ ⭐
- SECOND: Sparsely Embedded Convolutional Detection (Sensors 2018) Y. Yan et al. [pdf] [Github] ⭐ ⭐ ⭐
- PointPillars: Fast Encoders for Object Detection from Point Clouds (CVPR 2019), A. Lang et al. [pdf] [GIthub]
⭐ ⭐ ⭐ ⭐ ⭐ - Part-A^2 Net: 3D Part-Aware and Aggregation Neural Network for Object Detection from Point Cloud (ArXiv 2019) S. Shi et al. [pdf] [Github] ⭐ ⭐ ⭐ ⭐
Point-based methods
- PointRCNN: 3D Object Proposal Generation and Detection from Point Cloud (CVPR 2019), S. Shi et al. [pdf] [Github] ⭐ ⭐ ⭐ ⭐ ⭐
- Deep Hough Voting for 3D Object Detection in Point Clouds (ICCV 2019) C. R. Qi et al. [pdf] [Github]
⭐ ⭐ ⭐ ⭐ ⭐
Combining point-based and grid-based methods
- STD: Sparse-to-Dense 3D Object Detector for Point Cloud (ICCV 2019), Z. Yang et al. [pdf] ⭐ ⭐ ⭐ ⭐
- Fast Point R-CNN (ICCV 2019), Y. Chen et al. [pdf] ⭐ ⭐ ⭐
- PV-RCNN: Point-Voxel Feature Set Abstraction for 3D Object Detection (Arxiv 2019) S. Shi et al. [pdf]
⭐ ⭐ ⭐ ⭐ ⭐
- IPOD: Intensive Point-based Object Detector for Point Cloud (ArXiv 2018) Z. Yang et al. [pdf] ⭐ ⭐ ⭐ ⭐
- RoarNet: A Robust 3D Object Detection based on RegiOn Approximation Refinement ((ArXiv 2018), K. Shin et al. [pdf]
- Frustum PointNets for 3D Object Detection from RGB-D Data (CVPR 2018), C. R. Qi et al. [pdf] [GIthub]
⭐ ⭐ ⭐ ⭐ ⭐ - Frustum ConvNet: Sliding Frustums to Aggregate Local Point-Wise Features for Amodal 3D Object Detection (CVPR 2019), Z. Wang et al. [pdf] ⭐ ⭐ ⭐
- Multi-View 3D Object Detection Network for Autonomous Driving (CVPR 2017), X. Chen et al. [pdf] [Github]
⭐ ⭐ ⭐ ⭐ - PointFusion: Deep Sensor Fusion for 3D Bounding Box Estimation (CVPR 2018), D. Xu et al. [pdf] ⭐ ⭐ ⭐ ⭐
- Deep Continuous Fusion for Multi-Sensor 3D Object Detection (ECCV 2018), M. Liang et al. [pdf] ⭐ ⭐ ⭐ ⭐
- Multi-Task Multi-Sensor Fusion for 3D Object Detection (CVPR 2019), M. Liang et al. [pdf] ⭐ ⭐ ⭐ ⭐ ⭐
- MVX-Net: Multimodal VoxelNet for 3D Object Detection (ICRA 2019), V. A. Sindagi et al. [pdf] ⭐ ⭐ ⭐ ⭐
- Recurrent Slice Networks for 3D Segmentation of Point Clouds (CVPR 2018), Q. Huang et al. [pdf] [Github]
⭐ ⭐ ⭐ ⭐ - SGPN: Similarity Group Proposal Network for 3D Point Cloud Instance Segmentation (CVPR 2018), W. Wang et al. [pdf] [Github] ⭐ ⭐ ⭐ ⭐
- Associatively Segmenting Instances and Semantics in Point Clouds (CVPR 2019), X. Long et al. [pdf]
⭐ ⭐ ⭐ ⭐ ⭐
...(To be completed)
Note that some of these datasets don't provide point cloud data, which means you need some toolboxes to convert data from mesh or RGB-D images.