汇总3D点云语义分割论文（ALL）

参考来源：

1、https://github.com/Yochengliu/awesome-point-cloud-analysis

2、awesome-point-cloud-analysis-2021: A list of papers and datasets about point cloud analysis (processing) since 2017.

- Recent papers (from 2017)

Table of Contents

• 2017
• 2018
• 2019
• 2020 [CVPR: 70 papers; ECCV: 39 papers]
• 2021 [CVPR: 18/60 paper; ICCV: 10/39 papers]

1. [3D点云语义分割综述] from by [胡庆永](Deep Learning for 3D Point Clouds：A Survey_20200727版)
2. Yochengliu/awesome-point-cloud-analysis from 2017
3. NUAAXQ/awesome-point-cloud-analysis-2020
4. NUAAXQ/awesome-point-cloud-analysis-2021: A list of papers and datasets about point cloud analysis (processing) since 2017.

Keywords

dat.: dataset   |   cls.: classification   |   rel.: retrieval   |   seg.: segmentation
det.: detection   |   tra.: tracking   |   pos.: pose   |   dep.: depth
reg.: registration   |   rec.: reconstruction   |   aut.: autonomous driving
oth.: other, including normal-related, correspondence, mapping, matching, alignment, compression, generative model...

Statistics: 🔥 code is available & stars >= 100  |  ⭐ citation >= 50

2017

• [CVPR] PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation. [tensorflow][pytorch] [cls. seg. det.] 🔥 ⭐

• [CVPR] SyncSpecCNN: Synchronized Spectral CNN for 3D Shape Segmentation. [torch] [seg. oth.] ⭐

• [CVPR] ScanNet: Richly-annotated 3D Reconstructions of Indoor Scenes. [project][git] [dat. cls. rel. seg. oth.] 🔥 ⭐

• [CVPR] OctNet: Learning Deep 3D Representations at High Resolutions. [torch] [cls. seg. oth.] 🔥 ⭐

• [ICCV] Escape from Cells: Deep Kd-Networks for the Recognition of 3D Point Cloud Models. [pytorch] [cls. rel. seg.] ⭐

• [ICCV] 3DCNN-DQN-RNN: A Deep Reinforcement Learning Framework for Semantic Parsing of Large-scale 3D Point Clouds. [code] [seg.]

• [ICCV] 3D Graph Neural Networks for RGBD Semantic Segmentation. [pytorch] [seg.]

• [NeurIPS] PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Space. [tensorflow][pytorch] [cls. seg.] 🔥 ⭐

• [ICRA] Fast segmentation of 3D point clouds: A paradigm on LiDAR data for autonomous vehicle applications. [code] [seg. aut.]

• [ICRA] SegMatch: Segment based place recognition in 3D point clouds. [seg. oth.]

• [3DV] SEGCloud: Semantic Segmentation of 3D Point Clouds. [project] [seg. aut.] ⭐

2018

• [CVPR] SPLATNet: Sparse Lattice Networks for Point Cloud Processing. [caffe] [seg.] 🔥

• [CVPR] Attentional ShapeContextNet for Point Cloud Recognition. [cls. seg.]

• [CVPR] Mining Point Cloud Local Structures by Kernel Correlation and Graph Pooling. [code] [cls. seg.]

• [CVPR] Pointwise Convolutional Neural Networks. [tensorflow] [cls. seg.]

• [CVPR] SO-Net: Self-Organizing Network for Point Cloud Analysis. [pytorch] [cls. seg.] 🔥 ⭐

• [CVPR] Recurrent Slice Networks for 3D Segmentation of Point Clouds. [pytorch] [seg.]

• [CVPR] 3D Semantic Segmentation with Submanifold Sparse Convolutional Networks. [pytorch] [seg.] 🔥

• [CVPR] Deep Parametric Continuous Convolutional Neural Networks. [seg. aut.]

• [CVPR] SGPN: Similarity Group Proposal Network for 3D Point Cloud Instance Segmentation. [tensorflow] [seg.] 🔥

• [CVPR] Large-scale Point Cloud Semantic Segmentation with Superpoint Graphs. [pytorch] [seg.] 🔥

• [CVPR] Density Adaptive Point Set Registration. [code] [reg.]

• [CVPR] A Minimalist Approach to Type-Agnostic Detection of Quadrics in Point Clouds. [seg.]

• [CVPR] PointGrid: A Deep Network for 3D Shape Understanding. [tensorflow] [cls. seg.]

• [CVPR] Tangent Convolutions for Dense Prediction in 3D. [tensorflow] [seg. aut.]

• [ECCV] 3D Recurrent Neural Networks with Context Fusion for Point Cloud Semantic Segmentation. [seg.]

• [ECCV] Local Spectral Graph Convolution for Point Set Feature Learning. [tensorflow] [cls. seg.]

• [ECCV] SpiderCNN: Deep Learning on Point Sets with Parameterized Convolutional Filters. [tensorflow] [cls. seg.]

• [ECCV] Fully-Convolutional Point Networks for Large-Scale Point Clouds. [tensorflow] [seg. oth.]

• [ECCVW] 3DContextNet: K-d Tree Guided Hierarchical Learning of Point Clouds Using Local and Global Contextual Cues. [cls. seg.]

• [NeurIPS] PointCNN: Convolution On X-Transformed Points. [tensorflow][pytorch] [cls. seg.] 🔥

• [TOG] Point Convolutional Neural Networks by Extension Operators. [tensorflow] [cls. seg.]

• [SIGGRAPH Asia] Monte Carlo Convolution for Learning on Non-Uniformly Sampled Point Clouds. [tensorflow] [cls. seg. oth.]

• [SIGGRAPH] Learning local shape descriptors from part correspondences with multi-view convolutional networks. [project] [seg. oth.]

• [MM] RGCNN: Regularized Graph CNN for Point Cloud Segmentation. [tensorflow] [seg.]

• [ICRA] Multi-View 3D Entangled Forest for Semantic Segmentation and Mapping. [seg. oth.]

• [ICRA] SqueezeSeg: Convolutional Neural Nets with Recurrent CRF for Real-Time Road-Object Segmentation from 3D LiDAR Point Cloud. [tensorflow] [seg. aut.]

• [IROS] Extracting Phenotypic Characteristics of Corn Crops Through the Use of Reconstructed 3D Models. [seg. rec.]

• [ACCV] Flex-Convolution (Million-Scale Point-Cloud Learning Beyond Grid-Worlds). [tensorflow] [seg.]

---

2019

• [CVPR] Relation-Shape Convolutional Neural Network for Point Cloud Analysis. [pytorch] [cls. seg. oth.] 🔥
• [CVPR] Spherical Fractal Convolutional Neural Networks for Point Cloud Recognition. [cls. seg.]
• [CVPR] Modeling Point Clouds with Self-Attention and Gumbel Subset Sampling. [cls. seg.]
• [CVPR] A-CNN: Annularly Convolutional Neural Networks on Point Clouds. [tensorflow] [cls. seg.]
• [CVPR] PointConv: Deep Convolutional Networks on 3D Point Clouds. [tensorflow] [cls. seg.] 🔥
• [CVPR] Path-Invariant Map Networks. [tensorflow] [seg. oth.]
• [CVPR] PartNet: A Large-scale Benchmark for Fine-grained and Hierarchical Part-level 3D Object Understanding. [code] [dat. seg.]
• [CVPR] Associatively Segmenting Instances and Semantics in Point Clouds. [tensorflow] [seg.] 🔥
• [CVPR] Octree guided CNN with Spherical Kernels for 3D Point Clouds. [extension] [code] [cls. seg.]
• [CVPR] JSIS3D: Joint Semantic-Instance Segmentation of 3D Point Clouds with Multi-Task Pointwise Networks and Multi-Value Conditional Random Fields. [pytorch] [seg.]
• [CVPR] Point Cloud Oversegmentation with Graph-Structured Deep Metric Learning. [seg.]
• [CVPR] PartNet: A Recursive Part Decomposition Network for Fine-grained and Hierarchical Shape Segmentation. [pytorch] [dat. seg.]
• [CVPR] 4D Spatio-Temporal ConvNets: Minkowski Convolutional Neural Networks. [pytorch] [seg.] 🔥
• [CVPR] FilterReg: Robust and Efficient Probabilistic Point-Set Registration using Gaussian Filter and Twist Parameterization. [code] [reg.]
• [CVPR] PointWeb: Enhancing Local Neighborhood Features for Point Cloud Processing. [pytorch] [cls. seg.]
• [CVPR] GSPN: Generative Shape Proposal Network for 3D Instance Segmentation in Point Cloud. [seg.]
• [CVPR] Graph Attention Convolution for Point Cloud Semantic Segmentation. [seg.]
• [CVPR] LP-3DCNN: Unveiling Local Phase in 3D Convolutional Neural Networks. [project] [cls. seg.]
• [CVPR] Structural Relational Reasoning of Point Clouds. [cls. seg.]
• [ICCV] DeepGCNs: Can GCNs Go as Deep as CNNs? [tensorflow] [pytorch] [seg.] 🔥
• [ICCV] KPConv: Flexible and Deformable Convolution for Point Clouds. [tensorflow] [cls. seg.] 🔥
• [ICCV] ShellNet: Efficient Point Cloud Convolutional Neural Networks using Concentric Shells Statistics. [project] [seg.]
• [ICCV] DensePoint: Learning Densely Contextual Representation for Efficient Point Cloud Processing. [pytorch] [cls. seg. oth.]
• [ICCV] Hierarchical Point-Edge Interaction Network for Point Cloud Semantic Segmentation. [seg.]
• [ICCV] Interpolated Convolutional Networks for 3D Point Cloud Understanding. [cls. seg.]
• [ICCV] Dynamic Points Agglomeration for Hierarchical Point Sets Learning. [pytorch] [cls. seg.]
• [ICCV] Unsupervised Multi-Task Feature Learning on Point Clouds. [cls. seg.]
• [ICCV] VV-NET: Voxel VAE Net with Group Convolutions for Point Cloud Segmentation. [tensorflow] [seg.]
• [ICCV] MeteorNet: Deep Learning on Dynamic 3D Point Cloud Sequences. [code] [cls. seg. oth.]
• [ICCV] 3D Instance Segmentation via Multi-Task Metric Learning. [code] [seg.]
• [NeurIPS] Learning Object Bounding Boxes for 3D Instance Segmentation on Point Clouds. [tensorflow] [det. seg.]
• [NeurIPS] Exploiting Local and Global Structure for Point Cloud Semantic Segmentation with Contextual Point Representations. [tensorflow] [seg.]
• [NeurIPS] Point-Voxel CNN for Efficient 3D Deep Learning. [det. seg. aut.]
• [AAAI] Point2Sequence: Learning the Shape Representation of 3D Point Clouds with an Attention-based Sequence to Sequence Network. [tensorflow] [cls. seg.]
• [TOG] Dynamic Graph CNN for Learning on Point Clouds. [tensorflow][pytorch] [cls. seg.] 🔥 ⭐
• [SIGGRAPH Asia] RPM-Net: recurrent prediction of motion and parts from point cloud. [tensorflow] [seg.]
• [SIGGRAPH Asia] StructureNet: Hierarchical Graph Networks for 3D Shape Generation. [seg. oth.]
• [MM] SRINet: Learning Strictly Rotation-Invariant Representations for Point Cloud Classification and Segmentation. [tensorflow] [cls. seg.]
• [MM] Ground-Aware Point Cloud Semantic Segmentation for Autonomous Driving. [code] [seg. aut.]
• [ICRA] SqueezeSegV2: Improved Model Structure and Unsupervised Domain Adaptation for Road-Object Segmentation from a LiDAR Point Cloud. [tensorflow] [seg. aut.]
• [ICRA] Oriented Point Sampling for Plane Detection in Unorganized Point Clouds. [det. seg.]
• [ICRA] PointNetGPD: Detecting Grasp Configurations from Point Sets. [pytorch] [det. seg.]
• [ICRA] Robust 3D Object Classification by Combining Point Pair Features and Graph Convolution. [cls. seg.]
• [ICRA] Hierarchical Depthwise Graph Convolutional Neural Network for 3D Semantic Segmentation of Point Clouds. [seg.]
• [IROS] PASS3D: Precise and Accelerated Semantic Segmentation for 3D Point Cloud. [seg. aut.]
• [IV] End-to-End 3D-PointCloud Semantic Segmentation for Autonomous Driving. [seg.] [aut.]
• [Eurographics Workshop] Generalizing Discrete Convolutions for Unstructured Point Clouds. [pytorch] [cls. seg.]
• [3DV] Rotation Invariant Convolutions for 3D Point Clouds Deep Learning. [project] [cls. seg.]
• [3DV] Effective Rotation-invariant Point CNN with Spherical Harmonics kernels. [tensorflow] [cls. seg. oth.]

---

2020

• [AAAI] Point2Node: Correlation Learning of Dynamic-Node for Point Cloud Feature Modeling. [seg. cls.]

• [AAAI] PRIN: Pointwise Rotation-Invariant Network. [seg. cls.]

• [AAAI] JSNet: Joint Instance and Semantic Segmentation of 3D Point Clouds. [tensorflow][seg.][seg.]

• [CVPR] RandLA-Net: Efficient Semantic Segmentation of Large-Scale Point Clouds. [tensorflow] [seg.] 🔥

• [CVPR] Fusion-Aware Point Convolution for Online Semantic 3D Scene Segmentation. [pytorch] [seg.]

• [CVPR] AdaCoSeg: Adaptive Shape Co-Segmentation with Group Consistency Loss. [seg.]

• [CVPR] Point-GNN: Graph Neural Network for 3D Object Detection in a Point Cloud. [tensorflow][det.] 🔥

• [CVPR] Multi-Path Region Mining For Weakly Supervised 3D Semantic Segmentation on Point Clouds. [seg.]

• [CVPR] RPM-Net: Robust Point Matching using Learned Features. [code] [seg.]

• [CVPR] PolarNet: An Improved Grid Representation for Online LiDAR Point Clouds Semantic Segmentation. [pytorch] [seg.]

• [CVPR] 3D-MPA: Multi Proposal Aggregation for 3D Semantic Instance Segmentation. [seg.]

• [CVPR] OccuSeg: Occupancy-aware 3D Instance Segmentation. [seg.]

• [CVPR] Learning to Segment 3D Point Clouds in 2D Image Space. [pytorch] [seg]

• [CVPR] PointGroup: Dual-Set Point Grouping for 3D Instance Segmentation. [seg.]

• [CVPR] DualSDF: Semantic Shape Manipulation using a Two-Level Representation. [code] [seg]

• [CVPR] From Image Collections to Point Clouds with Self-supervised Shape and Pose Networks. [tensorflow] ['image-to-point cloud.']

• [CVPR] xMUDA: Cross-Modal Unsupervised Domain Adaptation for 3D Semantic Segmentation. [Segmentation]

• [CVPR] End-to-End 3D Point Cloud Instance Segmentation Without Detection. [Segmentation]

• [CVPR] Weakly Supervised Semantic Point Cloud Segmentation: Towards 10x Fewer Labels. [Segmentation]

• [CVPR] SegGCN: Efficient 3D Point Cloud Segmentation With Fuzzy Spherical Kernel. [Segmentation]

• [CVPR] SpSequenceNet: Semantic Segmentation Network on 4D Point Clouds. [Segmentation]

• [ECCV] PUGeo-Net: A Geometry-centric Network for 3D Point Cloud Upsampling. [Upsampling]

• [ECCV] Learning Graph-Convolutional Representations for Point Cloud Denoising. [Denoising]

• [ECCV] JSENet: Joint Semantic Segmentation and Edge Detection Network for 3D Point Clouds. [code] [Segmentation]

• [ECCV] Self-Prediction for Joint Instance and Semantic Segmentation of Point Clouds. [Segmentation]

• [ECCV] Virtual Multi-view Fusion for 3D Semantic Segmentation. [Segmentation]

• [ECCV] Searching Efficient 3D Architectures with Sparse Point-Voxel Convolution. [Segmentation]

• [ECCV] Rotation-robust Intersection over Union for 3D Object Detection. [3D IOU]

• [ECCV] Efficient Outdoor 3D Point Cloud Semantic Segmentation for Critical Road Objects and Distributed Contexts. [Segmentation]

• [ECCV] Deep FusionNet for Point Cloud Semantic Segmentation. [code] [Segmentation]

• [ECCV] SqueezeSegV3: Spatially-Adaptive Convolution for Efficient Point-Cloud Segmentation. [code] [seg.]

• [IROS] PBP-Net: Point Projection and Back-Projection Network for 3D Point Cloud Segmentation. [Segmentation ]

• [IROS] RegionNet: Region-feature-enhanced 3D Scene Understanding Network with Dual Spatial-aware Discriminative Loss. [Segmentation ]

• [IROS] Cascaded Non-local Neural Network for Point Cloud Semantic Segmentation. [Segmentation]

• [ACM MM] Campus3D: A Photogrammetry Point Cloud Benchmark for Hierarchical Understanding of Outdoor Scene. [Understanding]

• [WACV] FuseSeg: LiDAR Point Cloud Segmentation Fusing Multi-Modal Data. [seg. aut.]

• [WACV] Global Context Reasoning for Semantic Segmentation of 3D Point Clouds. [seg.]

• [BMVC] ASAP-Net: Attention and Structure Aware Point Cloud Sequence Segmentation. [Segmentation]

• [ICRA] DeepTemporalSeg: Temporally Consistent Semantic Segmentation of 3D LiDAR Scans. [seg.]

• [Master Thesis] Neighborhood Pooling in Graph Neural Networks for 3D and 4D Semantic Segmentation. ['seg.']

---

2021

• [CVPR] Towards Semantic Segmentation of Urban-Scale 3D Point Clouds: A Dataset, Benchmarks and Challenges. [code] [Segmentation]

• [CVPR] Monte Carlo Scene Search for 3D Scene Understanding. [Understanding]

• [CVPR] AF2-S3Net: Attentive Feature Fusion with Adaptive Feature Selection for Sparse Semantic Segmentation Network. [Segmentation]

• [CVPR] Equivariant Point Network for 3D Point Cloud Analysis. [Analysis]

• [CVPR] PAConv: Position Adaptive Convolution with Dynamic Kernel Assembling on Point Clouds. [code] [Convolution]

• [CVPR] Point Cloud Instance Segmentation using Probabilistic Embeddings. [Segmentation] ​ 💃

• [CVPR] Panoptic-PolarNet: Proposal-free LiDAR Point Cloud Panoptic Segmentation. [Segmentation]

• [CVPR oral] Point 4D Transformer Networks for Spatio-Temporal Modeling in Point Cloud Videos. [pytorch] [Transformer]

• [CVPR] Semantic Segmentation for Real Point Cloud Scenes via Bilateral Augmentation and Adaptive Fusion. [Segmentation]

• [CVPR] Complete & Label: A Domain Adaptation Approach to Semantic Segmentation of LiDAR Point Clouds. [Segmentation]

• [CVPR] PU-GCN: Point Cloud Upsampling using Graph Convolutional Networks. [code] [Upsampling]

• [CVPR] SCF-Net: Learning Spatial Contextual Features for Large-Scale Point Cloud Segmentation. [Segmentation] 💃

• [CVPR] Self-Supervised Learning on 3D Point Clouds by Learning Discrete Generative Models. [Self-Supervised]

• [CVPR] DyCo3D: Robust Instance Segmentation of 3D Point Clouds Through Dynamic Convolution. [code] [Segmentation]

• [CVPR] TearingNet: Point Cloud Autoencoder To Learn Topology-Friendly Representations. [Autoencoder]

• [CVPR] CGA-Net: Category Guided Aggregation for Point Cloud Semantic Segmentation. [Segmentation]

• [CVPR] Point Cloud Upsampling via Disentangled Refinement. [code] [Upsampling]

• [CVPR] Few-shot 3D Point Cloud Semantic Segmentation. [code] [Segmentation]

• [ICCV] DRINet: A Dual-Representation Iterative Learning Network for Point Cloud Segmentation. [Segmentation]

• [ICCV] ReDAL: Region-based and Diversity-aware Active Learning for Point Cloud Semantic Segmentation. [Segmentation]

• [ICCV] Learning with Noisy Labels for Robust Point Cloud Segmentation. [code][Segmentation]

• [ICCV] Instance Segmentation in 3D Scenes using Semantic Superpoint Tree Networks. [code][Segmentation]

• [ICCV] Hierarchical Aggregation for 3D Instance Segmentation. [code][Segmentation]

• [ICCV] Point Transformer. [Transformer]

• [ICCV] TempNet: Online Semantic Segmentation on Large-scale Point Cloud Series. [Segmentation]

• [ICCV] Perturbed Self-Distillation: Weakly Supervised Large-Scale Point Cloud Semantic Segmentation. [Segmentation]

• [ICCV] Guided Point Contrastive Learning for Semi-Supervised Point Cloud Semantic Segmentation. [Segmentation]

• [ICCV] Superpoint Network for Point Cloud Oversegmentation.[code] [Segmentation]

---

Datasets

• [ShapeNet] A collaborative dataset between researchers at Princeton, Stanford and TTIC. [seg.]
• [PartNet] The PartNet dataset provides fine grained part annotation of objects in ShapeNetCore. [seg.]
• [PartNet] PartNet benchmark from Nanjing University and National University of Defense Technology. [seg.]
• [S3DIS] The Stanford Large-Scale 3D Indoor Spaces Dataset. [seg.]
• [ScanNet] Richly-annotated 3D Reconstructions of Indoor Scenes. [cls. seg.]
• [UWA Dataset] . [cls. seg. reg.]
• [IQmulus & TerraMobilita Contest] The database contains 3D MLS data from a dense urban environment in Paris (France), composed of 300 million points. The acquisition was made in January 2013. [cls. seg. det.]
• [WAD] [ApolloScape] The datasets are provided by Baidu Inc. [tra. seg. det.]
• [SynthCity] SynthCity is a 367.9M point synthetic full colour Mobile Laser Scanning point cloud. Nine categories. [seg. aut.]
• [Lyft Level 5] Include high quality, human-labelled 3D bounding boxes of traffic agents, an underlying HD spatial semantic map. [det. seg. aut.]
• [SemanticKITTI] Sequential Semantic Segmentation, 28 classes, for autonomous driving. All sequences of KITTI odometry labeled. [ICCV 2019 paper] [seg. oth. aut.]
• [NPM3D] The Paris-Lille-3D has been produced by a Mobile Laser System (MLS) in two different cities in France (Paris and Lille). [seg.]
• [DALES] DALES: A Large-scale Aerial LiDAR Data Set for Semantic Segmentation. [seg.]