mlsanityData to 14 Oct 2026

Global Hierarchical Attention for 3D Point Cloud Analysis

arXiv 2208.03791 ↗First posted 7 Aug 2022cs.CVcited by 0

Not on the map. Why?

Abstract (arXiv)

We propose a new attention mechanism, called Global Hierarchical Attention (GHA), for 3D point cloud analysis. GHA approximates the regular global dot-product attention via a series of coarsening and interpolation operations over multiple hierarchy levels. The advantage of GHA is two-fold. First, it has linear complexity with respect to the number of points, enabling the processing of large point clouds. Second, GHA inherently possesses the inductive bias to focus on spatially close points, while retaining the global connectivity among all points. Combined with a feedforward network, GHA can be inserted into many existing network architectures. We experiment with multiple baseline networks and show that adding GHA consistently improves performance across different tasks and datasets. For the task of semantic segmentation, GHA gives a +1.7% mIoU increase to the MinkowskiEngine baseline on ScanNet. For the 3D object detection task, GHA improves the CenterPoint baseline by +0.5% mAP on the nuScenes dataset, and the 3DETR baseline by +2.1% mAP25 and +1.5% mAP50 on ScanNet.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 58 of its 73 references that are in the map. First posted on arXiv 7 Aug 2022.

Builds on

58 of 73 references.

References · 58

  1. 2017Attention Is All You Needcited by 60,361
  2. 2021Swin Transformer: Hierarchical Vision Transformer using Shifted Windowscited by 12,139
  3. 2020End-to-End Object Detection with Transformerscited by 7,414
  4. 2021Emerging Properties in Self-Supervised Vision Transformerscited by 6,968
  5. 2016PointNet: Deep Learning on Point Sets for 3D Classification and Segmentationcited by 5,248
  6. 2019nuScenes: A multimodal dataset for autonomous drivingcited by 4,523
  7. 2017PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Spacecited by 4,401
  8. 2017ScanNet: Richly-annotated 3D Reconstructions of Indoor Scenescited by 3,117
  9. 2018Dynamic Graph CNN for Learning on Point Cloudscited by 2,223
  10. 2020Fourier Features Let Networks Learn High Frequency Functions in Low Dimensional Domainscited by 2,200
  11. 2018PointPillars: Fast Encoders for Object Detection from Point Cloudscited by 1,715
  12. 2020Big Bird: Transformers for Longer Sequencescited by 1,684
38 more
  1. 2020Rethinking Attention with Performerscited by 1,601
  2. 2020Linformer: Self-Attention with Linear Complexitycited by 1,474
  3. 2020Point Transformercited by 1,470
  4. 2017VoxelNet: End-to-End Learning for Point Cloud Based 3D Object Detectioncited by 1,323
  5. 20194D Spatio-Temporal ConvNets: Minkowski Convolutional Neural Networkscited by 1,208
  6. 2019KPConv: Flexible and Deformable Convolution for Point Cloudscited by 1,205
  7. 2020Center-based 3D Object Detection and Trackingcited by 1,144
  8. 2018PointRCNN: 3D Object Proposal Generation and Detection from Point Cloudcited by 941
  9. 2019PV-RCNN: Point-Voxel Feature Set Abstraction for 3D Object Detectioncited by 891
  10. 2020Efficient Transformers: A Surveycited by 790
  11. 2018PointCNN: Convolution On X-Transformed Pointscited by 764
  12. 20173D Semantic Segmentation with Submanifold Sparse Convolutional Networkscited by 748
  13. 2020PCT: Point cloud transformercited by 732
  14. 2021Do Vision Transformers See Like Convolutional Neural Networks?cited by 675
  15. 2018PointConv: Deep Convolutional Networks on 3D Point Cloudscited by 657
  16. 2019Deep Hough Voting for 3D Object Detection in Point Cloudscited by 648
  17. 2017Frustum PointNets for 3D Object Detection from RGB-D Datacited by 640
  18. 20203DSSD: Point-based 3D Single Stage Object Detectorcited by 440
  19. 2017Large-scale Point Cloud Semantic Segmentation with Superpoint Graphscited by 385
  20. 2019From Points to Parts: 3D Object Detection from Point Cloud with Part-aware and Part-aggregation Networkcited by 368
  21. 2016OctNet: Learning Deep 3D Representations at High Resolutionscited by 342
  22. 20163DMatch: Learning Local Geometric Descriptors from RGB-D Reconstructionscited by 336
  23. 2021An End-to-End Transformer Model for 3D Object Detectioncited by 329
  24. 2019Deep Closest Point: Learning Representations for Point Cloud Registrationcited by 321
  25. 2020Exploring Self-attention for Image Recognitioncited by 313
  26. 2018SpiderCNN: Deep Learning on Point Sets with Parameterized Convolutional Filterscited by 261
  27. 2021Voxel Transformer for 3D Object Detectioncited by 247
  28. 2021Self-Supervised Pretraining of 3D Features on any Point-Cloudcited by 217
  29. 2021Group-Free 3D Object Detection via Transformerscited by 215
  30. 20203D Object Detection with Pointformercited by 194
  31. 2018Tangent Convolutions for Dense Prediction in 3Dcited by 173
  32. 2019End-to-End Multi-View Fusion for 3D Object Detection in LiDAR Point Cloudscited by 171
  33. 2018SPLATNet: Sparse Lattice Networks for Point Cloud Processingcited by 158
  34. 2021Mix3D: Out-of-Context Data Augmentation for 3D Scenescited by 126
  35. 2020H3DNet: 3D Object Detection Using Hybrid Geometric Primitivescited by 123
  36. 2021Fast Point Transformercited by 104
  37. 2021Bidirectional Projection Network for Cross Dimension Scene Understandingcited by 95
  38. 2019Interpolated Convolutional Networks for 3D Point Cloud Understandingcited by 86

50 most cited of 58 All 58 (JSON)

Built on by

Not cited yet.