Part Segmentation on ShapeNetPart (test)
87.5mIoU (Inst.)M2PT-Video
Evaluation results
| Method | Links | |||||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| M2PT-VideoTraining Protocol=Pretrained, Auxiliary Modality=Video2024.01 | 87.5 | 85.6 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| M2PT-ImageTraining Protocol=Pretrained, Auxiliary Modality=Image2024.01 | 87.5 | 85.6 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| M2PT-AudioTraining Protocol=Pretrained, Auxiliary Modality=Audio2024.01 | 87.5 | 85.6 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DeLAVoting=true2023.08 | 87.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 86 | |
| SPoTrYear=20232024.07 | 87.2 | 85.4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| SpoTr2024.03 | 87.2 | 85.4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| GSPoint2026.01 | 87.2 | 85.6 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| SPoTrVoting=true2023.08 | 87.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 85.4 | |
| PointNeXtYear=20222024.07 | 87.1 | 85.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| GSLCN (2023)2026.01 | 87.1 | 85.4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| PointNeXt-S (C=160)Voting=true, Channels (C)=1602023.08 | 87.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 85.4 | |
| DU-NetVoting=true2023.08 | 87.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 85.2 | |
| DeLAParams.=7.5M2024.03 | 87 | 85.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| PCMParams.=40.6M2024.03 | 87 | 85.3 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| PointNeXt (2022)2026.01 | 87 | 85.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DeLA2023.08 | 87 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 85.8 | |
| PointMLP+TAPPre-training Strategy=3D-to-2D Generative Pre-training, Backbone=PointMLP2023.07 | 86.9 | 85.2 | — | — | 84.8 | 86.1 | 89.5 | 82.5 | 92.1 | — | 92.3 | 88.7 | 85.6 | 96.5 | 79.8 | 96 | 85.9 | — | 78.1 | 83.2 | — | 75.9 | 66.2 | — | — | |
| PCM-TinyParams.=8.8M2024.03 | 86.9 | 85 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| PointNeXt-S (C=64)Voting=true, Channels (C)=642023.08 | 86.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 85.2 | |
| DC-CCNN++2026.06 | 86.9 | 85.2 | — | — | 84.1 | 83.5 | 89.8 | 81.1 | 92.4 | — | 92.7 | 88.2 | 83.4 | 96.3 | 77.3 | 96.1 | 86.2 | — | 83.5 | 83.6 | — | 77.9 | 68.4 | — | — | |
| CurveNetYear=20212024.07 | 86.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| GPSFormerYear=20242024.07 | 86.8 | 85.4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CurveNet2024.03 | 86.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| PointWavelet (2024)2026.01 | 86.8 | 85.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CurveNetVoting=true2023.08 | 86.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| ACT w/ LCMPretrain=MPM, Encoder=LCM2024.05 | 86.7 | 85 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| PointNeXt2025.04 | 86.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| PointMetaBase2025.04 | 86.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| SAMBLE2025.04 | 86.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Point-SRALearning Paradigm=Cross-Modal Self-Supervised Representation Learning2026.01 | 86.7 | 84.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| EDGCNet2025.12 | 86.7 | 84.4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| PointNeXt-SVoting=true2023.08 | 86.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 84.2 | |
| DU-Net2023.08 | 86.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 84.5 | |
| PointNeXt-STraining Paradigm=Supervised Learning Only2026.05 | 86.7 | 84.4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| PointTransformer2022.03 | 86.6 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Stratified Transformer2022.03 | 86.6 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| PointGPT-LProtocol=Self-Supervised Representation Learning, larger pre-training dataset=true, post-pre-training stage=true2023.05 | 86.6 | 84.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Point TransformerYear=20212024.07 | 86.6 | 83.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Point TransformerParams.=7.8M2024.03 | 86.6 | 83.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| PTv12024.03 | 86.6 | 83.7 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| MaskPoint w/ LCMPretrain=MPM, Encoder=LCM2024.05 | 86.6 | 85.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Point-MAE w/ LCMPretrain=MPM, Encoder=LCM2024.05 | 86.6 | 85.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| PT22025.04 | 86.6 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CurveNet2025.04 | 86.6 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DeltaConv2025.04 | 86.6 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Point Transformer2023.08 | 86.6 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 83.7 | |
| PVT2023.08 | 86.6 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| point-M2AEST=false2025.06 | 86.5 | 84.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| PointGPT-BProtocol=Self-Supervised Representation Learning, larger pre-training dataset=true, post-pre-training stage=true2023.05 | 86.5 | 84.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Point-M2AEPre-training Strategy=Uni-modal Generative Pre-training, Architecture=Transformer2023.07 | 86.5 | 84.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| PointNeXt-SC=160, Params.=22.5M2024.03 | 86.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Point-M2AEPretrain=MPM2024.05 | 86.5 | 84.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Point-BERT w/ LCMPretrain=MPM, Encoder=LCM2024.05 | 86.5 | 85 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Point-M2AE w/ LCMPretrain=MPM, Encoder=LCM2024.05 | 86.5 | 85 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DuGREAT (2025)2026.01 | 86.5 | 84.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| PointDico2025.12 | 86.5 | 84.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Point-M2AETraining Paradigm=Self-Supervised Pre-training with Single-Modal2026.05 | 86.5 | 84.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DC-CCNN2026.06 | 86.5 | 84.6 | — | — | 84.3 | 83 | 89.4 | 78.4 | 92.2 | — | 92.5 | 88.6 | 82.6 | 95.3 | 76.2 | 96.3 | 85.9 | — | 83.5 | 83.2 | — | 77.1 | 68.6 | — | — | |
| KPConv2021.03 | 86.4 | 85.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DensePoint2021.03 | 86.4 | 84.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| DensePoint2021.12 | 86.4 | 84.2 | — | — | 84 | 85.4 | 90 | 79.2 | 91.1 | 81.6 | 91.5 | 87.5 | 84.7 | 95.9 | 74.3 | 94.6 | 82.9 | 64.6 | 76.8 | 83.7 | — | — | — | — | — | |
| DensePoint2022.03 | 86.4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| ReConReference=ICML 2023, Protocol=Methods using cross-modal information and teacher models2023.05 | 86.4 | 84.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| RECON2023.02 | 86.4 | 84.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| KPConvPre-training Strategy=Supervised Representation Learning Only2023.07 | 86.4 | 85.1 | — | — | 84.6 | 86.3 | 87.2 | 81.1 | 91.1 | — | 92.6 | 88.4 | 82.7 | 96.2 | 78.1 | 95.8 | 85.4 | — | 82 | 83.6 | — | 77.8 | 69 | — | — | |
| KPConvYear=20192024.07 | 86.4 | 85.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| PCTYear=20212024.07 | 86.4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| AdaptConvYear=20212024.07 | 86.4 | 83.4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| ReConSelf-Supervised Representation Learning Mode=Cross-Modal2024.08 | 86.4 | 84.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| PointGPT-BPretrain=MPM2024.05 | 86.4 | 84.5 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| PCT2025.04 | 86.4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| ReConJournal or Conference=ICML, Learning Paradigm=Cross-Modal Self-Supervised Representation Learning2026.01 | 86.4 | 84.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| ReConLearning Protocol=Cross-modal / Teacher models2025.12 | 86.4 | 84.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| ReConEvaluation Protocol=Full Fine-tuning, #TP (M)=48.52026.05 | 86.4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | 84.5 | |
| ReConTraining Paradigm=Self-Supervised Pre-training with Cross-Modal2026.05 | 86.4 | 84.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| point-FEMAEST=false2025.06 | 86.3 | 84.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| InterpCNN2021.03 | 86.3 | 84 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| PRA-Net2021.12 | 86.3 | 83.7 | — | — | 84.4 | 86.8 | 89.5 | 78.4 | 91.4 | 76.4 | 91.5 | 88.2 | 85.3 | 95.7 | 73.4 | 94.8 | 82.1 | 62.3 | 75.5 | 84 | — | — | — | — | — | |
| InterpCNN2022.03 | 86.3 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Point-FEMAESelf-Supervised Representation Learning Mode=Single-Modal2024.08 | 86.3 | 84.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| LCMPretrain=None2024.05 | 86.3 | 84.6 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| PRA-Net2025.04 | 86.3 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Point-FEMAELearning Protocol=Self-Supervised (ALL)2025.12 | 86.3 | 84.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| Point-FEMAETraining Paradigm=Self-Supervised Pre-training with Single-Modal2026.05 | 86.3 | 84.9 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| PointGPT-SST=true2025.06 | 86.2 | 84.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| PVCNN2021.03 | 86.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| RS-CNNVoting=true2021.03 | 86.2 | 84 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| RSCNN2021.12 | 86.2 | 84 | — | — | 83.5 | 84.8 | 88.8 | 79.6 | 91.2 | 81.1 | 91.6 | 88.4 | 86 | 96 | 73.7 | 94.1 | 83.4 | 60.5 | 77.7 | 83.6 | — | — | — | — | — | |
| PVCNN2022.03 | 86.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| RS-CNN2022.03 | 86.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| KPConv2022.03 | 86.2 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| PointGPT-SProtocol=Self-Supervised Representation Learning2023.05 | 86.2 | 84.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| RS-CNNYear=20192024.07 | 86.2 | 84 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| PointGPTSelf-Supervised Representation Learning Mode=Single-Modal2024.08 | 86.2 | 84.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| PointGPT-SPretrain=MPM2024.05 | 86.2 | 84.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| PointMambaJournal or Conference=NeurIPS, Learning Paradigm=Self-Supervised Representation Learning2026.01 | 86.2 | 84.3 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| PointGPTLearning Protocol=Self-Supervised (ALL)2025.12 | 86.2 | 84.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| CSConLearning Protocol=Self-Supervised (ALL)2025.12 | 86.2 | 84.8 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| PointGPT-STraining Paradigm=Self-Supervised Pre-training with Single-Modal2026.05 | 86.2 | 84.1 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | |
| PointMambaTraining Paradigm=Self-Supervised Pre-training with Single-Modal2026.05 | 86.2 | 84.4 | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — | — |