mlsanityData to 14 Oct 2026

Revisiting Point Cloud Shape Classification with a Simple and Effective Baseline

arXiv 2106.05304 ↗First posted 9 Jun 2021cs.CV cs.LGcited by 145

Not on the map. Why?

Abstract (arXiv)

Processing point cloud data is an important component of many real-world systems. As such, a wide variety of point-based approaches have been proposed, reporting steady benchmark improvements over time. We study the key ingredients of this progress and uncover two critical results. First, we find that auxiliary factors like different evaluation schemes, data augmentation strategies, and loss functions, which are independent of the model architecture, make a large difference in performance. The differences are large enough that they obscure the effect of architecture. When these factors are controlled for, PointNet++, a relatively older network, performs competitively with recent methods. Second, a very simple projection-based method, which we refer to as SimpleView, performs surprisingly well. It achieves on par or better results than sophisticated state-of-the-art methods on ModelNet40 while being half the size of PointNet++. It also outperforms state-of-the-art methods on ScanObjectNN, a real-world point cloud benchmark, and demonstrates better cross-dataset generalization. Code is available at https://github.com/princeton-vl/SimpleView.

Key facts. Cited by 145 papers in the map (data through 14 Oct 2026). Builds on 43 of its 71 references that are in the map. First posted on arXiv 9 Jun 2021.

Builds on

43 of 71 references.

References · 43

  1. 2014Adam: A Method for Stochastic Optimizationcited by 37,240
  2. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 30,186
  3. 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,305
  4. 2015Rethinking the Inception Architecture for Computer Visioncited by 5,475
  5. 2016PointNet: Deep Learning on Point Sets for 3D Classification and Segmentationcited by 5,248
  6. 2017PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Spacecited by 4,401
  7. 2018Dynamic Graph CNN for Learning on Point Cloudscited by 2,223
  8. 2017Deep Setscited by 1,421
  9. 2019KPConv: Flexible and Deformable Convolution for Point Cloudscited by 1,205
  10. 2018PointCNN: Convolution On X-Transformed Pointscited by 764
  11. 2018PointConv: Deep Convolutional Networks on 3D Point Cloudscited by 657
  12. 2015Multi-view Convolutional Neural Networks for 3D Shape Recognitioncited by 565
31 more
  1. 2019Deep Learning for 3D Point Clouds: A Surveycited by 558
  2. 2019Revisiting Point Cloud Classification: A New Benchmark Dataset and Classification Model on Real-World Datacited by 532
  3. 2018Hypergraph Neural Networkscited by 488
  4. 2019Relation-Shape Convolutional Neural Network for Point Cloud Analysiscited by 346
  5. 2019Neural Point-Based Graphicscited by 327
  6. 2017Dynamic Edge-Conditioned Filters in Convolutional Neural Networks on Graphscited by 310
  7. 2018SpiderCNN: Deep Learning on Point Sets with Parameterized Convolutional Filterscited by 261
  8. 20196-DOF GraspNet: Variational Grasp Generation for Object Manipulationcited by 242
  9. 2016Volumetric and Multi-View CNNs for Object Classification on 3D Datacited by 240
  10. 2020PointASNL: Robust Point Clouds Processing using Nonlocal Neural Networks with Adaptive Samplingcited by 239
  11. 2019Billion-scale semi-supervised learning for image classificationcited by 232
  12. 2018SO-Net: Self-Organizing Network for Point Cloud Analysiscited by 210
  13. 2017SEGCloud: Semantic Segmentation of 3D Point Cloudscited by 179
  14. 2018Tangent Convolutions for Dense Prediction in 3Dcited by 173
  15. 2017Semantic3D.net: A new Large-scale Point Cloud Classification Benchmarkcited by 158
  16. 2018SPLATNet: Sparse Lattice Networks for Point Cloud Processingcited by 158
  17. 2018Point Convolutional Neural Networks by Extension Operatorscited by 154
  18. 2017Escape from Cells: Deep Kd-Networks for the Recognition of 3D Point Cloud Modelscited by 142
  19. 20183DMV: Joint 3D-Multi-View Prediction for 3D Semantic Scene Segmentationcited by 118
  20. 2015Deep Sliding Shapes for Amodal 3D Object Detection in RGB-D Imagescited by 111
  21. 2019DensePoint: Learning Densely Contextual Representation for Efficient Point Cloud Processingcited by 104
  22. 2018Point2Sequence: Learning the Shape Representation of 3D Point Clouds with an Attention-based Sequence to Sequence Networkcited by 100
  23. 2018Local Spectral Graph Convolution for Point Set Feature Learningcited by 84
  24. 2017Deep Projective 3D Semantic Segmentationcited by 81
  25. 2019Revealing Scenes by Inverting Structure from Motion Reconstructionscited by 66
  26. 2019A-CNN: Annularly Convolutional Neural Networks on Point Cloudscited by 66
  27. 20163D Shape Segmentation with Projective Convolutional Networkscited by 63
  28. 2016Analysis and Observations from the First Amazon Picking Challengecited by 41
  29. 2019Linked Dynamic Graph CNN: Learning on Point Cloud via Linking Hierarchical Featurescited by 31
  30. 2018Flex-Convolution (Million-Scale Point-Cloud Learning Beyond Grid-Worlds)cited by 23
  31. 2018Learning 3D Shapes as Multi-Layered Height-maps using 2D Convolutional Networkscited by 9

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2021: 9; 2022: 39; 2023: 36; 2024: 34; 2025: 19; 2026: 8.

145 papers.

Sort:

By area:

  1. 2021PointCLIP: Point Cloud Understanding by CLIPcited by 463
  2. 2022PointNeXt: Revisiting PointNet++ with Improved Training and Scaling Strategiescited by 436
  3. 2022Rethinking Network Design and Local Geometry in Point Cloud: A Simple Residual MLP Frameworkcited by 379
  4. 2023OmniObject3D: Large-Vocabulary 3D Object Dataset for Realistic Perception, Reconstruction and Generationcited by 326
  5. 2022ULIP: Learning a Unified Representation of Language, Images, and Point Clouds for 3D Understandingcited by 267
  6. 2023RVT: Robotic View Transformer for 3D Object Manipulationcited by 225
  7. 2022PointCLIP V2: Prompting CLIP and GPT for Powerful 3D Open-world Learningcited by 187
  8. 2023Contrast with Reconstruct: Contrastive 3D Representation Learning Guided by Generative Pretrainingcited by 164
  9. 2022CLIP2Point: Transfer CLIP to Point Cloud Classification with Image-Depth Pre-trainingcited by 160
  10. 2022Learning 3D Representations from 2D Pre-trained Models via Image-to-Point Masked Autoencoderscited by 144
  11. 2024RVT-2: Learning Precise Manipulation from Few Demonstrationscited by 137
  12. 2022Autoencoders as Cross-Modal Teachers: Can Pretrained 2D Image Transformers Help 3D Representation Learning?cited by 122
38 more
  1. 2022Benchmarking Robustness of 3D Point Cloud Recognition Against Common Corruptionscited by 80
  2. 2022Benchmarking and Analyzing Point Cloud Classification under Corruptionscited by 70
  3. 2023ViewRefer: Grasp the Multi-view Knowledge for 3D Visual Grounding with GPT and Prototype Guidancecited by 70
  4. 2023Instance-aware Dynamic Prompt Tuning for Pre-trained Point Cloud Modelscited by 68
  5. 2022LidarGait: Benchmarking 3D Gait Recognition with Point Cloudscited by 55
  6. 2023Joint-MAE: 2D-3D Joint Masked Autoencoders for 3D Point Cloud Pre-trainingcited by 52
  7. 2021Neural Point Light Fieldscited by 51
  8. 2022Shape-invariant 3D Adversarial Point Cloudscited by 50
  9. 2022A Comprehensive Survey on Deep Gait Recognition: Algorithms, Datasets and Challengescited by 45
  10. 2023Towards Compact 3D Representations via Point Feature Enhancement Masked Autoencoderscited by 42
  11. 2023Self-positioning Point-based Transformer for Point Cloud Understandingcited by 40
  12. 2023Frozen Transformers in Language Models Are Effective Visual Encoder Layerscited by 39
  13. 2023Parameter is Not All You Need: Starting from Non-Parametric Networks for 3D Point Cloud Analysiscited by 34
  14. 2024PointAD: Comprehending 3D Anomalies from Points and Pixels for Zero-shot 3D Anomaly Detectioncited by 29
  15. 2023Revisiting pre-trained remote sensing model benchmarks: resizing and normalization matterscited by 27
  16. 2022Isometric 3D Adversarial Examples in the Physical Worldcited by 25
  17. 2023Take-A-Photo: 3D-to-2D Generative Pre-training of Point Cloud Modelscited by 25
  18. 2022Let Images Give You More:Point Cloud Cross-Modal Training for Shape Analysiscited by 24
  19. 2024PCP-MAE: Learning to Predict Centers for Point Masked Autoencoderscited by 24
  20. 2022MATE: Masked Autoencoders are Online 3D Test-Time Learnerscited by 23
  21. 2021Voint Cloud: Multi-View Point Cloud Representation for 3D Understandingcited by 22
  22. 2022PointVector: A Vector Representation In Point Cloud Analysiscited by 21
  23. 2023Point Cloud Classification Using Content-based Transformer via Clustering in Feature Spacecited by 19
  24. 2022Rethinking the compositionality of point clouds through regularization in the hyperbolic spacecited by 18
  25. 2021DSPoint: Dual-scale Point Cloud Recognition with High-frequency Fusioncited by 17
  26. 2022Set Norm and Equivariant Skip Connections: Putting the Deep in Deep Setscited by 17
  27. 2022POS-BERT: Point Cloud One-Stage BERT Pre-Trainingcited by 16
  28. 2024CLIPose: Category-Level Object Pose Estimation with Pre-trained Vision-Language Knowledgecited by 16
  29. 2024Parameter-efficient Prompt Learning for 3D Point Cloud Understandingcited by 16
  30. 2024KPConvX: Modernizing Kernel Point Convolution with Kernel Attentioncited by 16
  31. 2024Bringing Masked Autoencoders Explicit Contrastive Properties for Point Cloud Self-Supervised Learningcited by 16
  32. 2024Unlearnable 3D Point Clouds: Class-wise Transformation Is All You Needcited by 16
  33. 2022Masked Surfel Prediction for Self-Supervised Point Cloud Learningcited by 15
  34. 2023A Large Scale Homography Benchmarkcited by 15
  35. 20243D Part Segmentation via Geometric Aggregation of 2D Visual Featurescited by 15
  36. 2022AGConv: Adaptive Graph Convolution on 3D Point Cloudscited by 13
  37. 2023Invariant Training 2D-3D Joint Hard Samples for Few-Shot Point Cloud Recognitioncited by 13
  38. 2021Interpreting Representation Quality of DNNs for 3D Point Cloud Processingcited by 12

50 most cited of 145 All 145 (JSON)