mlsanityData to 7 Oct 2026

Point Transformer V3: Simpler, Faster, Stronger

arXiv 2312.10035 ↗First posted 15 Dec 2023cs.CVcited by 579

Not on the map. Why?

Cited by 25 in Diffusion models, but placed by its own abstract, which didn’t fit.

Abstract (arXiv)

This paper is not motivated to seek innovation within the attention mechanism. Instead, it focuses on overcoming the existing trade-offs between accuracy and efficiency within the context of point cloud processing, leveraging the power of scale. Drawing inspiration from recent advances in 3D large-scale representation learning, we recognize that model performance is more influenced by scale than by intricate design. Therefore, we present Point Transformer V3 (PTv3), which prioritizes simplicity and efficiency over the accuracy of certain mechanisms that are minor to the overall performance after scaling, such as replacing the precise neighbor search by KNN with an efficient serialized neighbor mapping of point clouds organized with specific patterns. This principle enables significant scaling, expanding the receptive field from 16 to 1024 points while remaining efficient (a 3x increase in processing speed and a 10x improvement in memory efficiency compared with its predecessor, PTv2). PTv3 attains state-of-the-art results on over 20 downstream tasks that span both indoor and outdoor scenarios. Further enhanced with multi-dataset joint training, PTv3 pushes these results to a higher level.

Key facts. Cited by 579 papers in the map (data through 7 Oct 2026). Builds on 87 of its 106 references that are in the map. First posted on arXiv 15 Dec 2023.

Builds on

87 of 106 references.

References · 87

  1. 2017Attention Is All You Needcited by 59,932
  2. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 29,915
  3. 2015U-Net: Convolutional Networks for Biomedical Image Segmentationcited by 19,604
  4. 2023GPT-4 Technical Reportcited by 18,132
  5. 2023LLaMA: Open and Efficient Foundation Language Modelscited by 13,958
  6. 2021Swin Transformer: Hierarchical Vision Transformer using Shifted Windowscited by 12,082
  7. 2023Segment Anythingcited by 8,158
  8. 2021Emerging Properties in Self-Supervised Vision Transformerscited by 6,893
  9. 2020Scaling Laws for Neural Language Modelscited by 6,411
  10. 2016PointNet: Deep Learning on Point Sets for 3D Classification and Segmentationcited by 5,223
  11. 2019nuScenes: A multimodal dataset for autonomous drivingcited by 4,485
  12. 2017PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Spacecited by 4,384
38 more
  1. 2022FlashAttention: Fast and Memory-Efficient Exact Attention with IO-Awarenesscited by 3,611
  2. 2017ScanNet: Richly-annotated 3D Reconstructions of Indoor Scenescited by 3,088
  3. 2023FlashAttention-2: Faster Attention with Better Parallelism and Work Partitioningcited by 2,551
  4. 2019Scalability in Perception for Autonomous Driving: Waymo Open Datasetcited by 2,436
  5. 2023Efficient Streaming Language Models with Attention Sinkscited by 2,115
  6. 2018PointPillars: Fast Encoders for Object Detection from Point Cloudscited by 1,707
  7. 2019Generating Long Sequences with Sparse Transformerscited by 1,523
  8. 2020Point Transformercited by 1,462
  9. 2021Swin Transformer V2: Scaling Up Capacity and Resolutioncited by 1,215
  10. 2019KPConv: Flexible and Deformable Convolution for Point Cloudscited by 1,202
  11. 20194D Spatio-Temporal ConvNets: Minkowski Convolutional Neural Networkscited by 1,201
  12. 2020Center-based 3D Object Detection and Trackingcited by 1,143
  13. 2020On Layer Normalization in the Transformer Architecturecited by 906
  14. 2018PointCNN: Convolution On X-Transformed Pointscited by 764
  15. 20173D Semantic Segmentation with Submanifold Sparse Convolutional Networkscited by 746
  16. 2020PCT: Point cloud transformercited by 730
  17. 2016Multi-View 3D Object Detection Network for Autonomous Drivingcited by 721
  18. 2018PointConv: Deep Convolutional Networks on 3D Point Cloudscited by 656
  19. 2021Point-BERT: Pre-training 3D Point Cloud Transformers with Masked Point Modelingcited by 647
  20. 2019RandLA-Net: Efficient Semantic Segmentation of Large-Scale Point Cloudscited by 584
  21. 2015Multi-view Convolutional Neural Networks for 3D Shape Recognitioncited by 564
  22. 2021CSWin Transformer: A General Vision Transformer Backbone with Cross-Shaped Windowscited by 526
  23. 2020PointContrast: Unsupervised Pre-training for 3D Point Cloud Understandingcited by 502
  24. 2016Semantic Scene Completion from a Single Depth Imagecited by 468
  25. 2022Masked Autoencoders for Point Cloud Self-supervised Learningcited by 456
  26. 2022PointNeXt: Revisiting PointNet++ with Improved Training and Scaling Strategiescited by 432
  27. 2020Searching Efficient 3D Architectures with Sparse Point-Voxel Convolutioncited by 422
  28. 2021Conditional Positional Encodings for Vision Transformerscited by 402
  29. 2020Cylindrical and Asymmetrical 3D Convolution Networks for LiDAR Segmentationcited by 388
  30. 2017Large-scale Point Cloud Semantic Segmentation with Superpoint Graphscited by 384
  31. 2022Point Transformer V2: Grouped Vector Attention and Partition-based Poolingcited by 384
  32. 2022Rethinking Network Design and Local Geometry in Point Cloud: A Simple Residual MLP Frameworkcited by 376
  33. 2019Deep Closest Point: Learning Representations for Point Cloud Registrationcited by 318
  34. 2020PointGroup: Dual-Set Point Grouping for 3D Instance Segmentationcited by 318
  35. 2022Images Speak in Images: A Generalist Painter for In-Context Visual Learningcited by 317
  36. 2022Language-Grounded Indoor 3D Semantic Segmentation in the Wildcited by 243
  37. 2020PointASNL: Robust Point Clouds Processing using Nonlocal Neural Networks with Adaptive Samplingcited by 238
  38. 2021PAConv: Position Adaptive Convolution with Dynamic Kernel Assembling on Point Cloudscited by 211

50 most cited of 87 All 87 (JSON)

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2024: 83; 2025: 248; 2026: 248.

579 papers.

Sort:

By area:

  1. 2024Aligning Cyber Space with Physical World: A Comprehensive Survey on Embodied AIcited by 175
  2. 2024PointMamba: A Simple State Space Model for Point Cloud Analysiscited by 166
  3. 2025Sonata: Self-Supervised Learning of Reliable Point Representationscited by 108
  4. 2025GPT4Scene: Understand 3D Scenes from Videos with Vision-Language Modelscited by 106
  5. 2024SAMPart3D: Segment Any Part in 3D Objectscited by 95
  6. 2025BridgeVLA: Input-Output Alignment for Efficient 3D Manipulation Learning with Vision-Language Modelscited by 77
  7. 2026PointWorld: Scaling 3D World Models for In-The-Wild Robotic Manipulationcited by 74
  8. 2024Large Spatial Model: End-to-end Unposed Images to Semantic 3Dcited by 73
  9. 2024Point Mamba: A Novel Point Cloud Backbone Based on State Space Model with Octree-Based Ordering Strategycited by 71
  10. 2024Point Cloud Mamba: Point Cloud Learning via State Space Modelcited by 66
  11. 2024Voxel Mamba: Group-Free State Space Models for Point Cloud based 3D Object Detectioncited by 66
  12. 2025SpatialLM: Training Large Language Models for Structured Indoor Modelingcited by 63
38 more
  1. 2024Mamba3D: Enhancing Local Features for 3D Point Cloud Analysis via State Space Modelcited by 57
  2. 2025SceneSplat: Gaussian Splatting-based Scene Understanding with Vision-Language Pretrainingcited by 51
  3. 2025Dr. Splat: Directly Referring 3D Gaussian Splatting via Direct Language Embedding Registrationcited by 50
  4. 2024TEXGen: a Generative Diffusion Model for Mesh Texturescited by 46
  5. 2024OA-CNNs: Omni-Adaptive Sparse CNNs for 3D Semantic Segmentationcited by 43
  6. 2024Generative Gaussian Splatting for Unbounded 3D City Generationcited by 43
  7. 2024Parameter-Efficient Fine-Tuning in Spectral Domain for Point Cloud Learningcited by 43
  8. 2025Concerto: Joint 2D-3D Self-Supervised Learning Emerges Spatial Representationscited by 42
  9. 2025P3-SAM: Native 3D Part Segmentationcited by 39
  10. 2024Towards Generalizable Vision-Language Robotic Manipulation: A Benchmark and LLM-guided 3D Policycited by 37
  11. 2025Training-Free Efficient Video Generation via Dynamic Token Carvingcited by 33
  12. 2025GraspGen: A Diffusion-based Framework for 6-DOF Grasping with On-Generator Trainingcited by 32
  13. 2025DINO in the Room: Leveraging 2D Foundation Models for 3D Segmentationcited by 31
  14. 2024Find Any Part in 3Dcited by 30
  15. 2024PO3AD: Predicting Point Offsets toward Better 3D Point Cloud Anomaly Detectioncited by 30
  16. 2025MM-OR: A Large Multimodal Operating Room Dataset for Semantic Understanding of High-Intensity Surgical Environmentscited by 28
  17. 2025Spatial Understanding from Videos: Structured Prompts Meet Simulation Datacited by 28
  18. 2025OctGPT: Octree-based Multiscale Autoregressive Models for 3D Shape Generationcited by 27
  19. 2024GroupContrast: Semantic-aware Self-supervised Representation Learning for 3D Understandingcited by 26
  20. 2024RadarPillars: Efficient Object Detection from 4D Radar Point Cloudscited by 23
  21. 2024PCP-MAE: Learning to Predict Centers for Point Masked Autoencoderscited by 23
  22. 2025Locate 3D: Real-World Object Localization via Self-Supervised Learning in 3Dcited by 23
  23. 2024Flow4D: Leveraging 4D Voxel Network for LiDAR Scene Flow Estimationcited by 22
  24. 2024An End-to-End Robust Point Cloud Semantic Segmentation Network with Single-Step Conditional Diffusion Modelscited by 22
  25. 2025Particle-Grid Neural Dynamics for Learning Deformable Object Models from RGB-D Videoscited by 22
  26. 2024PointRWKV: Efficient RWKV-Like Model for Hierarchical Point Cloud Learningcited by 21
  27. 2025Erwin: A Tree-based Hierarchical Transformer for Large-scale Physical Systemscited by 21
  28. 2025Dream2Flow: Bridging Video Generation and Open-World Manipulation with 3D Object Flowcited by 21
  29. 2024Multimodality Helps Few-shot 3D Point Cloud Semantic Segmentationcited by 20
  30. 2024Generative Densification: Learning to Densify Gaussians for High-Fidelity Generalizable 3D Reconstructioncited by 18
  31. 2025Reasoning in Space via Grounding in the Worldcited by 18
  32. 2026Utonia: Toward One Encoder for All Point Cloudscited by 18
  33. 2024PLGS: Robust Panoptic Lifting with 3D Gaussian Splattingcited by 17
  34. 2025Generalized Few-shot 3D Point Cloud Segmentation with Vision-Language Modelcited by 17
  35. 2025LitePT: Lighter Yet Stronger Point Transformercited by 17
  36. 2024Unlearnable 3D Point Clouds: Class-wise Transformation Is All You Needcited by 16
  37. 2025AMB3R: Accurate Feed-forward Metric-scale 3D Reconstruction with Backendcited by 16
  38. 2024G3PT: Unleash the power of Autoregressive Modeling in 3D Generation via Cross-scale Querying Transformercited by 15

50 most cited of 579 All 579 (JSON)