mlsanityData to 7 Oct 2026

PointNeXt: Revisiting PointNet++ with Improved Training and Scaling Strategies

arXiv 2206.04670 ↗First posted 9 Jun 2022cs.CV cs.AIcited by 432

On the map: World models → World modeling from LiDAR generation

Abstract (arXiv)

PointNet++ is one of the most influential neural architectures for point cloud understanding. Although the accuracy of PointNet++ has been largely surpassed by recent networks such as PointMLP and Point Transformer, we find that a large portion of the performance gain is due to improved training strategies, i.e. data augmentation and optimization techniques, and increased model sizes rather than architectural innovations. Thus, the full potential of PointNet++ has yet to be explored. In this work, we revisit the classical PointNet++ through a systematic study of model training and scaling strategies, and offer two major contributions. First, we propose a set of improved training strategies that significantly improve PointNet++ performance. For example, we show that, without any change in architecture, the overall accuracy (OA) of PointNet++ on ScanObjectNN object classification can be raised from 77.9% to 86.1%, even outperforming state-of-the-art PointMLP. Second, we introduce an inverted residual bottleneck design and separable MLPs into PointNet++ to enable efficient and effective model scaling and propose PointNeXt, the next version of PointNets. PointNeXt can be flexibly scaled up and outperforms state-of-the-art methods on both 3D classification and segmentation tasks. For classification, PointNeXt reaches an overall accuracy of 87.7 on ScanObjectNN, surpassing PointMLP by 2.3%, while being 10x faster in inference. For semantic segmentation, PointNeXt establishes a new state-of-the-art performance with 74.9% mean IoU on S3DIS (6-fold cross-validation), being superior to the recent Point Transformer. The code and models are available at https://github.com/guochengqian/pointnext.

Key facts. Cited by 432 papers in the map (data through 7 Oct 2026). Placed in World modeling from LiDAR generation, one of 11 directions in World models. Builds on 53 of its 61 references that are in the map. First posted on arXiv 9 Jun 2022.

Builds on

53 of 61 references.

Closest ancestors · 2

Same direction, newest first

  1. 2017PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Space7 Jun
  2. 2016PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation2 Dec

Other references · 51

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2014Adam: A Method for Stochastic Optimizationcited by 36,960
  3. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 29,915
  4. 2015U-Net: Convolutional Networks for Biomedical Image Segmentationcited by 19,604
  5. 2021Swin Transformer: Hierarchical Vision Transformer using Shifted Windowscited by 12,082
  6. 2019EfficientNet: Rethinking Model Scaling for Convolutional Neural Networkscited by 6,415
  7. 2015Rethinking the Inception Architecture for Computer Visioncited by 5,451
  8. 2018MobileNetV2: Inverted Residuals and Linear Bottleneckscited by 4,773
  9. 2014Going Deeper with Convolutionscited by 4,231
  10. 2017MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applicationscited by 3,827
  11. 2022A ConvNet for the 2020scited by 3,545
  12. 2017ScanNet: Richly-annotated 3D Reconstructions of Indoor Scenescited by 3,088
38 more
  1. 2015ShapeNet: An Information-Rich 3D Model Repositorycited by 2,880
  2. 2018Dynamic Graph CNN for Learning on Point Cloudscited by 2,220
  3. 2020Point Transformercited by 1,462
  4. 2019KPConv: Flexible and Deformable Convolution for Point Cloudscited by 1,202
  5. 20194D Spatio-Temporal ConvNets: Minkowski Convolutional Neural Networkscited by 1,201
  6. 2021Scaling Vision Transformerscited by 993
  7. 2018PointCNN: Convolution On X-Transformed Pointscited by 764
  8. 20173D Semantic Segmentation with Submanifold Sparse Convolutional Networkscited by 746
  9. 2020PCT: Point cloud transformercited by 730
  10. 2018PointConv: Deep Convolutional Networks on 3D Point Cloudscited by 656
  11. 2021Point-BERT: Pre-training 3D Point Cloud Transformers with Masked Point Modelingcited by 647
  12. 2019RandLA-Net: Efficient Semantic Segmentation of Large-Scale Point Cloudscited by 584
  13. 2015Multi-view Convolutional Neural Networks for 3D Shape Recognitioncited by 564
  14. 2019Revisiting Point Cloud Classification: A New Benchmark Dataset and Classification Model on Real-World Datacited by 532
  15. 2019DeepGCNs: Can GCNs Go as Deep as CNNs?cited by 527
  16. 2022Rethinking Network Design and Local Geometry in Point Cloud: A Simple Residual MLP Frameworkcited by 376
  17. 2021ResNet strikes back: An improved training procedure in timmcited by 365
  18. 2019Relation-Shape Convolutional Neural Network for Point Cloud Analysiscited by 346
  19. 2019Point-Voxel CNN for Efficient 3D Deep Learningcited by 343
  20. 2018SpiderCNN: Deep Learning on Point Sets with Parameterized Convolutional Filterscited by 261
  21. 2018Tangent Convolutions for Dense Prediction in 3Dcited by 173
  22. 2021Walk in the Cloud: Learning Curves for Point Clouds Shape Analysiscited by 172
  23. 2021Training Graph Neural Networks with 1000 Layerscited by 169
  24. 2022Stratified Transformer for 3D Point Cloud Segmentationcited by 166
  25. 2021Revisiting ResNets: Improved Training and Scaling Strategiescited by 158
  26. 2021Revisiting Point Cloud Shape Classification with a Simple and Effective Baselinecited by 145
  27. 2020MVTN: Multi-View Transformation Network for 3D Shape Recognitioncited by 140
  28. 2021Fast Point Transformercited by 103
  29. 2019PU-GCN: Point Cloud Upsampling using Graph Convolutional Networkscited by 100
  30. 2020A Closer Look at Local Aggregation Operators in Point Cloud Analysiscited by 87
  31. 2022Surface Representation for Point Cloudscited by 84
  32. 2019DeepGCNs: Making GCNs Go as Deep as CNNscited by 72
  33. 2021Semantic Segmentation for Real Point Cloud Scenes via Bilateral Augmentation and Adaptive Fusioncited by 71
  34. 2022PolyLoss: A Polynomial Expansion Perspective of Classification Loss Functionscited by 62
  35. 2021PRA-Net: Point Relation-Aware Network for 3D Point Cloud Analysiscited by 58
  36. 2022Contrastive Boundary Learning for Point Cloud Segmentationcited by 56
  37. 2020Dense-Resolution Network for Point Cloud Classification and Segmentationcited by 45
  38. 2021PatchFormer: An Efficient Point Transformer with Patch Attentioncited by 43

50 most cited of 51 All 51 (JSON)

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2022: 17; 2023: 97; 2024: 127; 2025: 120; 2026: 71.

432 papers.

Sort:

By area:

  1. 2023Point Transformer V3: Simpler, Faster, Strongercited by 579
  2. 20243D Diffuser Actor: Policy Diffusion with 3D Scene Representationscited by 278
  3. 2022ULIP: Learning a Unified Representation of Language, Images, and Point Clouds for 3D Understandingcited by 262
  4. 2023ULIP-2: Towards Scalable Multimodal Pre-training for 3D Understandingcited by 199
  5. 2023Point-Bind & Point-LLM: Aligning Point Cloud with Multi-modality for 3D Understanding, Generation, and Instruction Followingcited by 192
  6. 2023OpenShape: Scaling Up 3D Shape Representation Towards Open-World Understandingcited by 183
  7. 2024PointMamba: A Simple State Space Model for Point Cloud Analysiscited by 166
  8. 2023Contrast with Reconstruct: Contrastive 3D Representation Learning Guided by Generative Pretrainingcited by 162
  9. 2024Understanding World or Predicting Future? A Comprehensive Survey of World Modelscited by 161
  10. 2023Meta-Transformer: A Unified Framework for Multimodal Learningcited by 136
  11. 2024ShapeLLM: Universal 3D Object Understanding for Embodied Interactioncited by 133
  12. 2022PartSLIP: Low-Shot Part Segmentation for 3D Point Clouds via Pretrained Image-Language Modelscited by 117
38 more
  1. 2023PointGPT: Auto-regressively Generative Pre-training from Point Cloudscited by 115
  2. 20243D Diffusion Policy: Generalizable Visuomotor Policy Learning via Simple 3D Representationscited by 115
  3. 2023Swin3D: A Pretrained Transformer Backbone for 3D Indoor Scene Understandingcited by 111
  4. 2025Sonata: Self-Supervised Learning of Reliable Point Representationscited by 108
  5. 2024ManiGaussian: Dynamic Gaussian Splatting for Multi-task Robotic Manipulationcited by 106
  6. 2023OneFormer3D: One Transformer for Unified Point Cloud Segmentationcited by 95
  7. 2024SAMPart3D: Segment Any Part in 3D Objectscited by 95
  8. 2025PARTFIELD: Learning 3D Feature Fields for Part Segmentation and Beyondcited by 91
  9. 2023Towards Large-scale 3D Representation Learning with Multi-dataset Point Prompt Trainingcited by 85
  10. 2025Vision-Language-Action Models for Robotics: A Review Towards Real-World Applicationscited by 82
  11. 2024General Flow as Foundation Affordance for Scalable Robot Learningcited by 80
  12. 2025BridgeVLA: Input-Output Alignment for Efficient 3D Manipulation Learning with Vision-Language Modelscited by 77
  13. 2022P2P: Tuning Pre-trained Image Models for Point Cloud Analysis with Point-to-Pixel Promptingcited by 74
  14. 2024Point Cloud Mamba: Point Cloud Learning via State Space Modelcited by 66
  15. 2023Visual Programming for Zero-shot Open-Vocabulary 3D Visual Groundingcited by 64
  16. 2025OmniPart: Part-Aware 3D Generation with Semantic Decoupling and Structural Cohesioncited by 60
  17. 2023Radar-Camera Fusion for Object Detection and Semantic Segmentation in Autonomous Driving: A Comprehensive Reviewcited by 57
  18. 2024Mamba3D: Enhancing Local Features for 3D Point Cloud Analysis via State Space Modelcited by 57
  19. 2025HoloPart: Generative 3D Part Amodal Segmentationcited by 57
  20. 2023Efficient 3D Semantic Segmentation with Superpoint Transformercited by 56
  21. 2023PolarNet: 3D Point Clouds for Language-Guided Robotic Manipulationcited by 56
  22. 2023PonderV2: Pave the Way for 3D Foundation Model with A Universal Pre-training Paradigmcited by 56
  23. 2023PillarNeXt: Rethinking Network Designs for 3D Object Detection in LiDAR Point Cloudscited by 54
  24. 2023Using a Waffle Iron for Automotive Point Cloud Semantic Segmentationcited by 53
  25. 2024Dynamic Adapter Meets Prompt Tuning: Parameter-Efficient Transfer Learning for Point Cloud Analysiscited by 52
  26. 2025Dr. Splat: Directly Referring 3D Gaussian Splatting via Direct Language Embedding Registrationcited by 50
  27. 2024GenSim2: Scaling Robot Data Generation with Multi-modal and Reasoning LLMscited by 48
  28. 2023Point Cloud Pre-training with Diffusion Modelscited by 47
  29. 2025SoFar: Language-Grounded Orientation Bridges Spatial Reasoning and Object Manipulationcited by 46
  30. 2024Lift3D Foundation Policy: Lifting 2D Large-Scale Pretrained Models for Robust 3D Robotic Manipulationcited by 45
  31. 2024OA-CNNs: Omni-Adaptive Sparse CNNs for 3D Semantic Segmentationcited by 43
  32. 2024Parameter-Efficient Fine-Tuning in Spectral Domain for Point Cloud Learningcited by 43
  33. 2023Self-positioning Point-based Transformer for Point Cloud Understandingcited by 40
  34. 2023Explore In-Context Learning for 3D Point Cloud Understandingcited by 40
  35. 2022Meta Architecture for Point Cloud Analysiscited by 39
  36. 2023PartSLIP++: Enhancing Low-Shot 3D Part Segmentation via Multi-View Instance Segmentation and Maximum Likelihood Estimationcited by 37
  37. 2023Parameter is Not All You Need: Starting from Non-Parametric Networks for 3D Point Cloud Analysiscited by 34
  38. 2024SegPoint: Segment Any Point Cloud via Large Language Modelcited by 34

50 most cited of 432 All 432 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2024PointInfinity: Resolution-Invariant Point Diffusion Models
  2. 2024DAPoinTr: Domain Adaptive Point Transformer for Point Cloud Completion
  3. 2025PointDGRWKV: Generalizing RWKV-like Architecture to Unseen Domains for Point Cloud Classification
  4. 2025SODA: Out-of-Distribution Detection in Domain-Shifted Point Clouds via Neighborhood Propagation
  5. 2023DiffPMAE: Diffusion Masked Autoencoders for Point Cloud Reconstruction
  6. 2018Nesti-Net: Normal Estimation for Unstructured 3D Point Clouds using Convolutional Neural Networks
  7. 2025PointLAMA: Latent Attention meets Mamba for Efficient Point Cloud Pretraining
  8. 2026Deformation-based In-Context Learning for Point Cloud Understanding