TasksSotADatasetsPapersMethodsSubmitAbout
Papers With Code 2

A community resource for machine learning research: papers, code, benchmarks, and state-of-the-art results.

Explore

Notable BenchmarksAll SotADatasetsPapersMethods

Community

Submit ResultsAbout

Data sourced from the PWC Archive (CC-BY-SA 4.0). Built by the community, for the community.

Models/PointNeXt

PointNeXt

Reported on 31 benchmarks across 5 tasks · 1 paper · 9 SOTA

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Computer Vision21 results

  • Shape Representation Of 3D Point CloudsonScanObjectNN
    Mean Accuracy· 2022-06-09
    86.8
    best: 93.8 (GPSFormer)
    SOTA
    PointNeXt: Revisiting PointNet++ with Improved Training and Scaling StrategiesarXiv:2206.04670
  • Shape Representation Of 3D Point CloudsonScanObjectNN
    Overall Accuracy· 2022-06-09
    88.2
    best: 97.2 (OmniVec2)
    SOTA
    PointNeXt: Revisiting PointNet++ with Improved Training and Scaling StrategiesarXiv:2206.04670
  • Shape Representation Of 3D Point CloudsonScanObjectNN
    Overall Accuracy (PB_T50_RS)· 2022-06-09
    87.8
    best: 92.64 (Mamba3D)
    SOTA
    PointNeXt: Revisiting PointNet++ with Improved Training and Scaling StrategiesarXiv:2206.04670
  • 3D Point Cloud ClassificationonScanObjectNN
    Mean Accuracy· 2022-06-09
    86.8
    best: 93.8 (GPSFormer)
    SOTA
    PointNeXt: Revisiting PointNet++ with Improved Training and Scaling StrategiesarXiv:2206.04670
  • 3D Point Cloud ClassificationonScanObjectNN
    Overall Accuracy· 2022-06-09
    88.2
    best: 97.2 (OmniVec2)
    SOTA
    PointNeXt: Revisiting PointNet++ with Improved Training and Scaling StrategiesarXiv:2206.04670
  • 3D Point Cloud ClassificationonScanObjectNN
    Overall Accuracy (PB_T50_RS)· 2022-06-09
    87.8
    best: 92.64 (Mamba3D)
    SOTA
    PointNeXt: Revisiting PointNet++ with Improved Training and Scaling StrategiesarXiv:2206.04670
  • 3D Point Cloud ReconstructiononScanObjectNN
    Mean Accuracy· 2022-06-09
    86.8
    best: 93.8 (GPSFormer)
    SOTA
    PointNeXt: Revisiting PointNet++ with Improved Training and Scaling StrategiesarXiv:2206.04670
  • 3D Point Cloud ReconstructiononScanObjectNN
    Overall Accuracy· 2022-06-09
    88.2
    best: 97.2 (OmniVec2)
    SOTA
    PointNeXt: Revisiting PointNet++ with Improved Training and Scaling StrategiesarXiv:2206.04670
  • 3D Point Cloud ReconstructiononScanObjectNN
    Overall Accuracy (PB_T50_RS)· 2022-06-09
    87.8
    best: 92.64 (Mamba3D)
    SOTA
    PointNeXt: Revisiting PointNet++ with Improved Training and Scaling StrategiesarXiv:2206.04670
  • Shape Representation Of 3D Point CloudsonModelNet40
    Mean Accuracy· 2022-06-09
    91.1
    best: 92.4 (ULIP + PointMLP)
    PointNeXt: Revisiting PointNet++ with Improved Training and Scaling StrategiesarXiv:2206.04670
  • Shape Representation Of 3D Point CloudsonModelNet40
    Overall Accuracy· 2022-06-09
    94
    best: 95.3 (PointGST)
    PointNeXt: Revisiting PointNet++ with Improved Training and Scaling StrategiesarXiv:2206.04670
  • Shape Representation Of 3D Point CloudsonScanObjectNN
    GFLOPs· 2022-06-09
    3.6
    best: 45 (PCM)
    PointNeXt: Revisiting PointNet++ with Improved Training and Scaling StrategiesarXiv:2206.04670
  • Shape Representation Of 3D Point CloudsonScanObjectNN
    Number of params (M)· 2022-06-09
    1.4
    best: 34.2 (PCM)
    PointNeXt: Revisiting PointNet++ with Improved Training and Scaling StrategiesarXiv:2206.04670
  • 3D Point Cloud ClassificationonModelNet40
    Mean Accuracy· 2022-06-09
    91.1
    best: 92.4 (ULIP + PointMLP)
    PointNeXt: Revisiting PointNet++ with Improved Training and Scaling StrategiesarXiv:2206.04670
  • 3D Point Cloud ClassificationonModelNet40
    Overall Accuracy· 2022-06-09
    94
    best: 95.3 (PointGST)
    PointNeXt: Revisiting PointNet++ with Improved Training and Scaling StrategiesarXiv:2206.04670
  • 3D Point Cloud ClassificationonScanObjectNN
    GFLOPs· 2022-06-09
    3.6
    best: 45 (PCM)
    PointNeXt: Revisiting PointNet++ with Improved Training and Scaling StrategiesarXiv:2206.04670
  • 3D Point Cloud ClassificationonScanObjectNN
    Number of params (M)· 2022-06-09
    1.4
    best: 34.2 (PCM)
    PointNeXt: Revisiting PointNet++ with Improved Training and Scaling StrategiesarXiv:2206.04670
  • 3D Point Cloud ReconstructiononModelNet40
    Mean Accuracy· 2022-06-09
    91.1
    best: 92.4 (ULIP + PointMLP)
    PointNeXt: Revisiting PointNet++ with Improved Training and Scaling StrategiesarXiv:2206.04670
  • 3D Point Cloud ReconstructiononModelNet40
    Overall Accuracy· 2022-06-09
    94
    best: 95.3 (PointGST)
    PointNeXt: Revisiting PointNet++ with Improved Training and Scaling StrategiesarXiv:2206.04670
  • 3D Point Cloud ReconstructiononScanObjectNN
    GFLOPs· 2022-06-09
    3.6
    best: 45 (PCM)
    PointNeXt: Revisiting PointNet++ with Improved Training and Scaling StrategiesarXiv:2206.04670
  • 3D Point Cloud ReconstructiononScanObjectNN
    Number of params (M)· 2022-06-09
    1.4
    best: 34.2 (PCM)
    PointNeXt: Revisiting PointNet++ with Improved Training and Scaling StrategiesarXiv:2206.04670

Medical5 results

  • Semantic SegmentationonS3DIS Area5
    mAcc· 2022-06-09
    77.2
    best: 81.6 (Sonata + PTv3)
    PointNeXt: Revisiting PointNet++ with Improved Training and Scaling StrategiesarXiv:2206.04670
  • Semantic SegmentationonS3DIS Area5
    mIoU· 2022-06-09
    71.1
    best: 76 (Sonata + PTv3)
    PointNeXt: Revisiting PointNet++ with Improved Training and Scaling StrategiesarXiv:2206.04670
  • Semantic SegmentationonS3DIS Area5
    oAcc· 2022-06-09
    91
    best: 93 (Sonata + PTv3)
    PointNeXt: Revisiting PointNet++ with Improved Training and Scaling StrategiesarXiv:2206.04670
  • Semantic SegmentationonShapeNet-Part
    Class Average IoU· 2022-06-09
    85.2
    best: 87.7 (Feature Geometric Net (FG-Net))
    PointNeXt: Revisiting PointNet++ with Improved Training and Scaling StrategiesarXiv:2206.04670
  • Semantic SegmentationonShapeNet-Part
    Instance Average IoU· 2022-06-09
    87.1
    best: 89.1 (GeomGCNN)
    PointNeXt: Revisiting PointNet++ with Improved Training and Scaling StrategiesarXiv:2206.04670

Audio5 results

  • 10-shot image generationonS3DIS Area5
    mAcc· 2022-06-09
    77.2
    best: 81.6 (Sonata + PTv3)
    PointNeXt: Revisiting PointNet++ with Improved Training and Scaling StrategiesarXiv:2206.04670
  • 10-shot image generationonS3DIS Area5
    mIoU· 2022-06-09
    71.1
    best: 76 (Sonata + PTv3)
    PointNeXt: Revisiting PointNet++ with Improved Training and Scaling StrategiesarXiv:2206.04670
  • 10-shot image generationonS3DIS Area5
    oAcc· 2022-06-09
    91
    best: 93 (Sonata + PTv3)
    PointNeXt: Revisiting PointNet++ with Improved Training and Scaling StrategiesarXiv:2206.04670
  • 10-shot image generationonShapeNet-Part
    Class Average IoU· 2022-06-09
    85.2
    best: 87.7 (Feature Geometric Net (FG-Net))
    PointNeXt: Revisiting PointNet++ with Improved Training and Scaling StrategiesarXiv:2206.04670
  • 10-shot image generationonShapeNet-Part
    Instance Average IoU· 2022-06-09
    87.1
    best: 89.1 (GeomGCNN)
    PointNeXt: Revisiting PointNet++ with Improved Training and Scaling StrategiesarXiv:2206.04670