Object Classification on ScanObjectNN PB_T50_RS
94.83AccuracyPointGST
Evaluation Results
| Method | Links | |
|---|---|---|
| PointGSTPre-trained model=PointGPT-L, Params. (M)=2.4, FLOPS (G)=67.952024.10 | 94.83 | |
| AsymDSD-B*Evaluation Protocol=Full Fine-tune, PP (Post-pretraining)=false, CM (Cross-modal training)=false, #P(M)=92.12025.06 | 93.72 | |
| Mamba3DPre-train=UniPre3D2025.06 | 93.4 | |
| Fully fine-tunePre-trained model=PointGPT-L, Params. (M)=360.5, FLOPS (G)=67.712024.10 | 93.4 | |
| ReCon++-B*Evaluation Protocol=Full Fine-tune, PP (Post-pretraining)=true, CM (Cross-modal training)=true, #P(M)=177.42025.06 | 93.34 | |
| Mamba3DPT=Point-MAE, #P (M)=16.9, #F (G)=3.9, Voting Strategy=true2024.04 | 93.34 | |
| DAPTPre-trained model=PointGPT-L, Params. (M)=4.2, FLOPS (G)=71.642024.10 | 93.02 | |
| IDPTPre-trained model=PointGPT-L, Params. (M)=10.0, FLOPS (G)=75.192024.10 | 92.99 | |
| AsymDSD-S*Evaluation Protocol=Full Fine-tune, PP (Post-pretraining)=false, CM (Cross-modal training)=false, #P(M)=22.12025.06 | 92.89 | |
| Point-PEFTPre-trained model=PointGPT-L, Params. (M)=3.1, FLOPS (G)=73.622024.10 | 92.85 | |
| Mamba3DPT=×, #P (M)=16.9, #F (G)=3.9, Voting Strategy=true2024.04 | 92.64 | |
| Mamba3DPre-train=None2025.06 | 92.6 | |
| Mamba3DPT=Point-MAE, #P (M)=16.9, #F (G)=3.9, Voting Strategy=false2024.04 | 92.05 | |
| AsymDSD-B*Evaluation Protocol=Linear, PP (Post-pretraining)=false, CM (Cross-modal training)=false, #P(M)=91.82025.06 | 91.96 | |
| PointGPT-B*Evaluation Protocol=Full Fine-tune, PP (Post-pretraining)=true, CM (Cross-modal training)=false, #P(M)=92.02025.06 | 91.9 | |
| Mamba3DPT=×, #P (M)=16.9, #F (G)=3.9, Voting Strategy=false2024.04 | 91.81 | |
| AsymDSD-S*Evaluation Protocol=Linear, PP (Post-pretraining)=false, CM (Cross-modal training)=false, #P(M)=21.82025.06 | 91.31 | |
| PointGPT-LEvaluation Protocol=Full Fine-tune, PP (Post-pretraining)=false, CM (Cross-modal training)=false, #P(M)=310.02025.06 | 91.1 | |
| Point-SRAParadigm=Cross-Modal Self-Supervised, Params (M)=40.12026.01 | 90.77 | |
| ReConParadigm=Cross-Modal Self-Supervised, Params (M)=44.32026.01 | 90.63 | |
| AsymDSD-S#P(M)=22.1, ST (Standard Transformer)=true, SM (Single-modal training)=true, Evaluation Protocol=Full Fine-tune2025.06 | 90.53 | |
| DeLAPT=×, #P (M)=5.3, #F (G)=1.52024.04 | 90.4 | |
| PointConTPT=×2024.04 | 90.3 | |
| Point-RAE#P(M)=29.2, ST (Standard Transformer)=false, SM (Single-modal training)=true, Evaluation Protocol=Full Fine-tune2025.06 | 90.28 | |
| Point-FEMAE#P(M)=27.4, ST (Standard Transformer)=false, SM (Single-modal training)=true, Evaluation Protocol=Full Fine-tune2025.06 | 90.22 | |
| Point-FEMAEParadigm=Single-Modal Self-Supervised, Params (M)=41.52026.01 | 90.22 | |
| Point-MAE-B*Evaluation Protocol=Full Fine-tune, PP (Post-pretraining)=true, CM (Cross-modal training)=false, #P(M)=120.12025.06 | 90.2 | |
| Mamba3DPT=Point-BERT, #P (M)=16.9, #F (G)=3.9, Voting Strategy=false2024.04 | 90.11 | |
| I2P-MAEParadigm=Cross-Modal Self-Supervised, Params (M)=15.32026.01 | 90.11 | |
| Fully fine-tunePre-trained model=RECON, Params. (M)=22.1, FLOPS (G)=4.762024.10 | 90.01 | |
| ReCon SM#P(M)=43.6, ST (Standard Transformer)=false, SM (Single-modal training)=true, Evaluation Protocol=Full Fine-tune2025.06 | 89.73 | |
| RECONPublication=ICML 23, Tunable Params.=43.6 M, Evaluation Protocol=Self-Supervised Representation Learning (Full Fine-Tuning)2025.04 | 89.73 | |
| PointGPT-BEvaluation Protocol=Full Fine-tune, PP (Post-pretraining)=false, CM (Cross-modal training)=false, #P(M)=92.02025.06 | 89.6 | |
| PointMLPPre-train=UniPre3D2025.06 | 89.5 | |
| PointGSTPre-trained model=RECON, Params. (M)=0.6, FLOPS (G)=4.812024.10 | 89.49 | |
| DAPTPre-trained model=RECON, Params. (M)=1.1, FLOPS (G)=4.962024.10 | 89.38 | |
| PointMamba#P(M)=12.3, ST (Standard Transformer)=false, SM (Single-modal training)=true, Evaluation Protocol=Full Fine-tune2025.06 | 89.31 | |
| PointMambaParadigm=Single-Modal Self-Supervised, Params (M)=12.32026.01 | 89.31 | |
| P2P-HorNetPT=✓, #F (G)=34.62024.04 | 89.3 | |
| P2PParadigm=Supervised Learning, Params (M)=195.82026.01 | 89.3 | |
| PointNTPPretrain Framework=fully causal autoregression, Auxiliary Module=none2026.05 | 89.3 | |
| PCMPre-train=UniPre3D2025.06 | 89 | |
| AsymSD-CLS-S#P(M)=22.1, ST (Standard Transformer)=true, SM (Single-modal training)=true, Evaluation Protocol=Full Fine-tune2025.06 | 88.72 | |
| SPoTrPT=×, #P (M)=1.7, #F (G)=10.82024.04 | 88.6 | |
| AsymSD-MPM-S#P(M)=22.1, ST (Standard Transformer)=true, SM (Single-modal training)=true, Evaluation Protocol=Full Fine-tune2025.06 | 88.58 | |
| Point-MAE† w/ IDPT#TP (M)=1.7, Learning Paradigm=Self-Supervised Representation Learning (IDPT)2023.04 | 88.51 | |
| PointMLPPre-train=TAP [60]2025.06 | 88.5 | |
| Point-MAE†#TP (M)=22.1, Learning Paradigm=Self-Supervised Representation Learning (Full Fine-tuning)2023.04 | 88.41 | |
| PointGSTPre-trained model=ACT, Params. (M)=0.6, FLOPS (G)=4.812024.10 | 88.27 | |
| Fully fine-tunePre-trained model=ACT, Params. (M)=22.1, FLOPS (G)=4.762024.10 | 88.21 | |
| ACT#TP (M)=22.1, Learning Paradigm=Self-Supervised Representation Learning (Full Fine-tuning)2023.04 | 88.21 | |
| ACTPublication=ICLR'23, Tunable Params.=22.1 M, Evaluation Protocol=Self-Supervised Representation Learning (Full Fine-Tuning)2025.04 | 88.21 | |
| ACTParadigm=Cross-Modal Self-Supervised, Params (M)=22.12026.01 | 88.21 | |
| ACTPretrain Framework=cross-modal teacher-guided masked modeling, Auxiliary Module=teacher autoencoder + transformer decoder2026.05 | 88.21 | |
| IAE (M2AE)Pretrain Framework=implicit autoencoder, Auxiliary Module=implicit decoder2026.05 | 88.2 | |
| PCMPre-train=None2025.06 | 88.1 | |
| Standard TransformerPre-train=UniPre3D2025.06 | 87.93 | |
| UniPre3D (Std. Transformer)Pretrain Framework=cross-modal Gaussian splatting pretraining, Auxiliary Module=Gaussian predictor + fusion block2026.05 | 87.93 | |
| PointMetaBase-SParam (M)=0.62026.01 | 87.9 | |
| PointNeXt#P(M)=1.4, ST (Standard Transformer)=false, SM (Single-modal training)=true, Evaluation Protocol=Supervised2025.06 | 87.7 | |
| PointNeXtPublication=NeurIPS'22, Tunable Params.=1.4 M, Evaluation Protocol=Traditional Supervised Learning Only2025.04 | 87.7 | |
| PointNeXt-SParam (M)=1.52026.01 | 87.7 | |
| PointNeXtParadigm=Supervised Learning, Params (M)=1.42026.01 | 87.7 | |
| PointNeXtPretrain Framework=supervised from scratch, Auxiliary Module=N/A2026.05 | 87.7 | |
| IDPTPre-trained model=ACT, Params. (M)=1.7, FLOPS (G)=7.102024.10 | 87.65 | |
| ACT w/ IDPT#TP (M)=1.7, Learning Paradigm=Self-Supervised Representation Learning (IDPT)2023.04 | 87.65 | |
| Standard TransformerPre-train=PointDif [82]2025.06 | 87.61 | |
| PointDifParadigm=Single-Modal Self-Supervised2026.01 | 87.61 | |
| PointDifPretrain Framework=diffusion-based pretraining, Auxiliary Module=conditional point generator2026.05 | 87.61 | |
| Point-JEPAParadigm=Single-Modal Self-Supervised2026.01 | 87.6 | |
| DAPTPre-trained model=ACT, Params. (M)=1.1, FLOPS (G)=4.962024.10 | 87.54 | |
| ADSPublication=ICCV'23, Evaluation Protocol=Traditional Supervised Learning Only2025.04 | 87.5 | |
| ADSPretrain Framework=supervised from scratch, Auxiliary Module=N/A2026.05 | 87.5 | |
| PointMLPPre-train=None2025.06 | 87.4 | |
| P2P-RN101Pretrain Framework=supervised from scratch, Auxiliary Module=N/A2026.05 | 87.4 | |
| IDPTPre-trained model=RECON, Params. (M)=1.7, FLOPS (G)=7.102024.10 | 87.27 | |
| PointGPT-S#P(M)=22.1, ST (Standard Transformer)=true, SM (Single-modal training)=true, Evaluation Protocol=Full Fine-tune2025.06 | 86.9 | |
| PointGPT-SPretrain Framework=autoregressive generation, Auxiliary Module=extractor-generator transformer decoder2026.05 | 86.9 | |
| Point-M2AE#P(M)=15.3, ST (Standard Transformer)=false, SM (Single-modal training)=true, Evaluation Protocol=Full Fine-tune2025.06 | 86.43 | |
| Point-M2AELearning Paradigm=Self-Supervised Representation Learning (Full Fine-tuning)2023.04 | 86.43 | |
| Point-M2AEPT=Point-M2AE, #P (M)=15.3, #F (G)=3.62024.04 | 86.43 | |
| Point-M2AEPublication=NeurIPS 22, Tunable Params.=15.3 M, Evaluation Protocol=Self-Supervised Representation Learning (Full Fine-Tuning)2025.04 | 86.43 | |
| Point-M2AEParadigm=Single-Modal Self-Supervised, Params (M)=15.32026.01 | 86.43 | |
| Point-M2AEPretrain Framework=masked autoencoder, Auxiliary Module=transformer decoder2026.05 | 86.43 | |
| Point-GNParam (M)=02026.01 | 86.4 | |
| Point-PEFTPre-trained model=RECON, Params. (M)=0.7, FLOPS (G)=7.612024.10 | 86.36 | |
| Joint-MAEPretrain Framework=joint masked autoencoding, Auxiliary Module=transformer decoder2026.05 | 86.07 | |
| Standard TransformerPre-train=Point-CMAE [44]2025.06 | 85.95 | |
| Point-PEFTPre-trained model=ACT, Params. (M)=0.7, FLOPS (G)=7.612024.10 | 85.74 | |
| PointMLPLearning Paradigm=Supervised Learning Only2023.04 | 85.7 | |
| Standard TransformerPre-train=TAP [60]2025.06 | 85.67 | |
| TAPPretrain Framework=cross-modal generative pre-training, Auxiliary Module=2D generator2026.05 | 85.67 | |
| PointGSTPre-trained model=Point-BERT, Params. (M)=0.6, FLOPS (G)=4.812024.10 | 85.64 | |
| Point-MAE + Point LoRAPublication=Ours, Tunable Params.=0.77 M (3.43%), Evaluation Protocol=Self-Supervised Representation Learning (Parameter-Efficient Fine-Tuning)2025.04 | 85.53 | |
| DAPTPre-trained model=Point-BERT, Params. (M)=1.1, FLOPS (G)=4.962024.10 | 85.43 | |
| PointMLP#P(M)=12.6, ST (Standard Transformer)=true, SM (Single-modal training)=true, Evaluation Protocol=Supervised2025.06 | 85.4 | |
| PointMLPPublication=ICLR'22, Tunable Params.=13.2 M, Evaluation Protocol=Traditional Supervised Learning Only2025.04 | 85.4 | |
| PointMLPParam (M)=12.62026.01 | 85.4 | |
| PointMLPPretrain Framework=supervised from scratch, Auxiliary Module=N/A2026.05 | 85.4 | |
| PointGSTPre-trained model=Point-MAE, Params. (M)=0.6, FLOPS (G)=4.812024.10 | 85.29 |