mlsanityData to 14 Oct 2026

LargeKernel3D: Scaling up Kernels in 3D Sparse CNNs

arXiv 2206.10555 ↗First posted 21 Jun 2022cs.CV cs.LGcited by 72

Not on the map. Why?

Abstract (arXiv)

Recent advance in 2D CNNs has revealed that large kernels are important. However, when directly applying large convolutional kernels in 3D CNNs, severe difficulties are met, where those successful module designs in 2D become surprisingly ineffective on 3D networks, including the popular depth-wise convolution. To address this vital challenge, we instead propose the spatial-wise partition convolution and its large-kernel module. As a result, it avoids the optimization and efficiency issues of naive 3D large kernels. Our large-kernel 3D CNN network, LargeKernel3D, yields notable improvement in 3D tasks of semantic segmentation and object detection. It achieves 73.9% mIoU on the ScanNetv2 semantic segmentation and 72.8% NDS nuScenes object detection benchmarks, ranking 1st on the nuScenes LIDAR leaderboard. The performance further boosts to 74.2% NDS with a simple multi-modal fusion. In addition, LargeKernel3D can be scaled to 17x17x17 kernel size on Waymo 3D object detection. For the first time, we show that large kernels are feasible and essential for 3D visual tasks.

Key facts. Cited by 72 papers in the map (data through 14 Oct 2026). Builds on 68 of its 81 references that are in the map. First posted on arXiv 21 Jun 2022.

Builds on

68 of 81 references.

References · 68

  1. 2017Attention Is All You Needcited by 60,361
  2. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 30,186
  3. 2014Microsoft COCO: Common Objects in Contextcited by 17,608
  4. 2019Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformercited by 14,599
  5. 2021Swin Transformer: Hierarchical Vision Transformer using Shifted Windowscited by 12,139
  6. 2016Layer Normalizationcited by 5,669
  7. 2016PointNet: Deep Learning on Point Sets for 3D Classification and Segmentationcited by 5,248
  8. 2019nuScenes: A multimodal dataset for autonomous drivingcited by 4,523
  9. 2017PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Spacecited by 4,401
  10. 2017MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applicationscited by 3,845
  11. 2022A ConvNet for the 2020scited by 3,579
  12. 2016Gaussian Error Linear Units (GELUs)cited by 3,502
38 more
  1. 2017ScanNet: Richly-annotated 3D Reconstructions of Indoor Scenescited by 3,117
  2. 2019XLNet: Generalized Autoregressive Pretraining for Language Understandingcited by 2,811
  3. 2019Scalability in Perception for Autonomous Driving: Waymo Open Datasetcited by 2,451
  4. 2017Rethinking Atrous Convolution for Semantic Image Segmentationcited by 2,161
  5. 2018PointPillars: Fast Encoders for Object Detection from Point Cloudscited by 1,715
  6. 2017Deformable Convolutional Networkscited by 1,488
  7. 2020Point Transformercited by 1,470
  8. 2021Swin Transformer V2: Scaling Up Capacity and Resolutioncited by 1,228
  9. 20194D Spatio-Temporal ConvNets: Minkowski Convolutional Neural Networkscited by 1,208
  10. 2019KPConv: Flexible and Deformable Convolution for Point Cloudscited by 1,205
  11. 2020Center-based 3D Object Detection and Trackingcited by 1,144
  12. 2015Multi-Scale Context Aggregation by Dilated Convolutionscited by 1,133
  13. 2018Self-Attention with Relative Position Representationscited by 1,120
  14. 2019PV-RCNN: Point-Voxel Feature Set Abstraction for 3D Object Detectioncited by 891
  15. 2018PointCNN: Convolution On X-Transformed Pointscited by 764
  16. 20173D Semantic Segmentation with Submanifold Sparse Convolutional Networkscited by 748
  17. 2021Do Vision Transformers See Like Convolutional Neural Networks?cited by 675
  18. 2018Deformable ConvNets v2: More Deformable, Better Resultscited by 670
  19. 2018PointConv: Deep Convolutional Networks on 3D Point Cloudscited by 657
  20. 2021CSWin Transformer: A General Vision Transformer Backbone with Cross-Shaped Windowscited by 527
  21. 2022TransFusion: Robust LiDAR-Camera Fusion for 3D Object Detection with Transformerscited by 466
  22. 20203DSSD: Point-based 3D Single Stage Object Detectorcited by 440
  23. 2020Searching Efficient 3D Architectures with Sparse Point-Voxel Convolutioncited by 423
  24. 2020Voxel R-CNN: Towards High Performance Voxel-based 3D Object Detectioncited by 410
  25. 2021Conditional Positional Encodings for Vision Transformerscited by 406
  26. 2019From Points to Parts: 3D Object Detection from Point Cloud with Part-aware and Part-aggregation Networkcited by 368
  27. 2019Attention Augmented Convolutional Networkscited by 363
  28. 2019Class-balanced Grouping and Sampling for Point Cloud 3D Object Detectioncited by 350
  29. 2021Focal Self-attention for Local-Global Interactions in Vision Transformerscited by 288
  30. 2021Global Filter Networks for Image Classificationcited by 260
  31. 2021Voxel Transformer for 3D Object Detectioncited by 247
  32. 2020PointASNL: Robust Point Clouds Processing using Nonlocal Neural Networks with Adaptive Samplingcited by 239
  33. 2017Large Kernel Matters -- Improve Semantic Segmentation by Global Convolutional Networkcited by 208
  34. 2021Rethinking and Improving Relative Position Encoding for Vision Transformercited by 207
  35. 2022Unifying Voxel-based Representation with Transformer for 3D Object Detectioncited by 202
  36. 2022Patches Are All You Need?cited by 201
  37. 2021Embracing Single Stride 3D Object Detector with Sparse Transformercited by 193
  38. 2022Stratified Transformer for 3D Point Cloud Segmentationcited by 167

50 most cited of 68 All 68 (JSON)

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2022: 1; 2023: 13; 2024: 27; 2025: 19; 2026: 12.

72 papers.

Sort:

By area:

  1. 2023Point Transformer V3: Simpler, Faster, Strongercited by 589
  2. 2025Sonata: Self-Supervised Learning of Reliable Point Representationscited by 110
  3. 2023Towards Large-scale 3D Representation Learning with Multi-dataset Point Prompt Trainingcited by 86
  4. 2023UniPAD: A Universal Pre-training Paradigm for Autonomous Drivingcited by 73
  5. 2023Cross Modal Transformer: Towards Fast and Robust 3D Object Detectioncited by 70
  6. 2024Voxel Mamba: Group-Free State Space Models for Point Cloud based 3D Object Detectioncited by 66
  7. 2023UniRepLKNet: A Universal Perception Large-Kernel ConvNet for Audio, Video, Point Cloud, Time-Series and Image Recognitioncited by 64
  8. 2023PonderV2: Pave the Way for 3D Foundation Model with A Universal Pre-training Paradigmcited by 56
  9. 2024SparseAD: Sparse Query-Centric Paradigm for Efficient End-to-End Autonomous Drivingcited by 47
  10. 2024LION: Linear Group RNN for 3D Object Detection in Point Cloudscited by 38
  11. 2023BEVNeXt: Reviving Dense BEV Frameworks for 3D Object Detectioncited by 35
  12. 2024GraphBEV: Towards Robust BEV Feature Alignment for Multi-Modal 3D Object Detectioncited by 33
38 more
  1. 2024HDRFlow: Real-Time HDR Video Reconstruction with Large Motionscited by 31
  2. 2023Fully Sparse Fusion for 3D Object Detectioncited by 29
  3. 2024PeLK: Parameter-efficient Large Kernel ConvNets with Peripheral Convolutioncited by 28
  4. 2022BEV-MAE: Bird's Eye View Masked Autoencoders for Point Cloud Pre-training in Autonomous Driving Scenarioscited by 21
  5. 2024DifFUSER: Diffusion Model for Robust Multi-Sensor Fusion in 3D Object Detection and BEV Segmentationcited by 18
  6. 2024MCTrack: A Unified 3D Multi-Object Tracking Framework for Autonomous Drivingcited by 17
  7. 2024ProtoOcc: Accurate, Efficient 3D Occupancy Prediction Using Dual Branch Encoder-Prototype Query Decodercited by 17
  8. 2025LitePT: Lighter Yet Stronger Point Transformercited by 17
  9. 2024KPConvX: Modernizing Kernel Point Convolution with Kernel Attentioncited by 16
  10. 2024GAFusion: Adaptive Fusing LiDAR and Camera with Multiple Guidance for 3D Object Detectioncited by 15
  11. 2024SFPNet: Sparse Focal Point Network for Semantic Segmentation on General LiDAR Point Cloudscited by 13
  12. 2023UniM2AE: Multi-modal Masked Autoencoders with Unified 3D Representation for 3D Perception in Autonomous Drivingcited by 12
  13. 2023Decoupled Local Aggregation for Point Cloud Learningcited by 12
  14. 2025SAMFusion: Sensor-Adaptive Multimodal Fusion for 3D Object Detection in Adverse Weathercited by 12
  15. 2024Progressive Multi-Modal Fusion for Robust 3D Object Detectioncited by 11
  16. 2023PillarNeSt: Embracing Backbone Scaling and Pretraining for Pillar-based 3D Object Detectioncited by 10
  17. 2025UniMamba: Unified Spatial-Channel Representation Learning with Group-Efficient Mamba for LiDAR-based 3D Object Detectioncited by 10
  18. 2025FSHNet: Fully Sparse Hybrid Network for 3D Object Detectioncited by 9
  19. 2024Revealing the Dark Secrets of Extremely Large Kernel ConvNets on Robustnesscited by 8
  20. 2024Scaling Up Your Kernels: Large Kernel Design in ConvNets towards Universal Representationscited by 8
  21. 2024CT3D++: Improving 3D Object Detection with Keypoint-induced Channel-wise Transformercited by 7
  22. 2024SparseDet: A Simple and Effective Framework for Fully Sparse LiDAR-based 3D Object Detectioncited by 7
  23. 2025BFANet: Revisiting 3D Semantic Segmentation with Boundary Feature Analysiscited by 5
  24. 2024RepNeXt: A Fast Multi-Scale CNN using Structural Reparameterizationcited by 4
  25. 2026AdaSFormer: Adaptive Serialized Transformers for Monocular Semantic Scene Completion from Indoor Environmentscited by 3
  26. 2023Minuet: Accelerating 3D Sparse Convolutions on GPUscited by 2
  27. 2024Brain Age Estimation with a Greedy Dual-Stream Model for Limited Datasetscited by 2
  28. 2024From Histopathology Images to Cell Clouds: Learning Slide Representations with Hierarchical Cell Transformercited by 2
  29. 2025CoreNet: Conflict Resolution Network for Point-Pixel Misalignment and Sub-Task Suppression of 3D LiDAR-Camera Object Detectioncited by 2
  30. 2025IMM-MOT: A Novel 3D Multi-object Tracking Framework with Interacting Multiple Model Filtercited by 2
  31. 2025MonoCT: Overcoming Monocular 3D Detection Domain Shift with Consistent Teacher Modelscited by 2
  32. 2026Offline-Poly: A Polyhedral Framework For Offline 3D Multi-Object Trackingcited by 2
  33. 2023Generalized Robot 3D Vision-Language Model with Fast Rendering and Pre-Training Vision-Language Alignmentcited by 1
  34. 2024Large receptive field strategy and important feature extraction strategy in 3D object detectioncited by 1
  35. 2024AVS-Net: Point Sampling with Adaptive Voxel Size for 3D Scene Understandingcited by 1
  36. 2024FFNet: MetaMixer-based Efficient Convolutional Mixer Designcited by 1
  37. 2024Efficient 3D Perception on Multi-Sweep Point Cloud with Gumbel Spatial Pruningcited by 1
  38. 2024SEGT: A General Spatial Expansion Group Transformer for nuScenes Lidar-based Object Detection Taskcited by 1

50 most cited of 72 All 72 (JSON)