Denoising on ShapeNet In-Context
2.2L1 CD ErrorACT
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| ACTModel training/learning paradigm=Task-specific models (trained separately), Venues=ICLR'232024.11 | 2.2 | 2.3 | 2.2 | 2.3 | 2.5 | 2.3 | — | |
| ACTTraining Strategy=Task-Specific2024.04 | 2.2 | 2.3 | 2.2 | 2.3 | 2.5 | 2.3 | — | |
| ACTTraining Paradigm=Task-Specific2026.04 | 2.2 | 2.3 | 2.2 | 2.3 | 2.5 | 2.3 | — | |
| PCTModel training/learning paradigm=Task-specific models (trained separately), Venues=CVM'212024.11 | 2.3 | 2.2 | 2.2 | 2.2 | 2.3 | 2.2 | — | |
| PCTTraining Strategy=Task-Specific2024.04 | 2.3 | 2.2 | 2.2 | 2.2 | 2.3 | 2.2 | — | |
| PCTTraining Paradigm=Task-Specific2026.04 | 2.3 | 2.2 | 2.2 | 2.2 | 2.3 | 2.2 | — | |
| DeformPICTraining Paradigm=In-Context Learning2026.04 | 2.8 | 3.3 | 3.6 | 3.8 | 3.9 | 3.5 | — | |
| PIC++Training Strategy=In-context Learning, Aggregation=-Cat2024.04 | 3.8 | 4.3 | 5.1 | 5.6 | 6.7 | 5.1 | — | |
| PIC-CatModel training/learning paradigm=In-context learning models, Venues=NeurIPS'232024.11 | 3.9 | 4.6 | 5.3 | 6 | 6.8 | 5.3 | — | |
| PICTraining Strategy=In-context Learning, Aggregation=-Cat2024.04 | 3.9 | 4.6 | 5.3 | 6 | 6.8 | 5.3 | — | |
| PIC-CatTraining Paradigm=In-Context Learning2026.04 | 3.9 | 4.6 | 5.3 | 6 | 6.8 | 5.3 | — | |
| PointNetModel training/learning paradigm=Task-specific models (trained separately), Venues=CVPR'172024.11 | 4.1 | 4 | 4.1 | 4 | 4.2 | 4.1 | — | |
| PointNetTraining Strategy=Task-Specific2024.04 | 4.1 | 4 | 4.1 | 4 | 4.2 | 4.1 | — | |
| PointNetTraining Paradigm=Task-Specific2026.04 | 4.1 | 4 | 4.1 | 4 | 4.2 | 4.1 | — | |
| PIC-Cat + MICASModel training/learning paradigm=In-context learning models, Venues=Ours2024.11 | 4.2 | 4.4 | 4.6 | 4.9 | 5.1 | 4.6 | — | |
| DGCNNModel training/learning paradigm=Task-specific models (trained separately), Venues=TOG' 192024.11 | 4.7 | 4.5 | 4.6 | 4.5 | 4.7 | 4.6 | — | |
| DGCNNTraining Strategy=Task-Specific2024.04 | 4.7 | 4.5 | 4.6 | 4.5 | 4.7 | 4.6 | — | |
| DGCNNTraining Paradigm=Task-Specific2026.04 | 4.7 | 4.5 | 4.6 | 4.5 | 4.7 | 4.6 | — | |
| PIC-S-CatTraining Paradigm=In-Context Learning2026.04 | 4.7 | 5.7 | 6.5 | 7.4 | 8.2 | 6.5 | — | |
| PIC-Sep + MICASModel training/learning paradigm=In-context learning models, Venues=Ours2024.11 | 4.9 | 5.2 | 5.5 | 5.7 | 5.1 | 5.1 | — | |
| Point-MAEModel training/learning paradigm=Multi-task models: share backbone + multi-task heads, Venues=ECCV'222024.11 | 5.6 | 5.4 | 5.6 | 5.5 | 5.8 | 5.6 | — | |
| PointMAE‡Training Strategy=Multitask‡2024.04 | 5.6 | 5.4 | 5.6 | 5.5 | 5.8 | 5.6 | — | |
| Point-MAETraining Paradigm=Multitask, Pre-trained=Yes2026.04 | 5.6 | 5.4 | 5.6 | 5.5 | 5.8 | 5.6 | — | |
| PIC-S-CatModel training/learning paradigm=In-context learning models, Venues=Arxiv'242024.11 | 5.7 | 5.7 | 6.5 | 7.4 | 8.2 | 6.7 | — | |
| PIC++Training Strategy=In-context Learning, Aggregation=-Sep2024.04 | 5.8 | 7.1 | 7.7 | 8.2 | 8.4 | 7.4 | — | |
| PICTraining Strategy=In-context Learning, Aggregation=-Sep2024.04 | 6.3 | 7.2 | 7.9 | 8.2 | 8.6 | 7.6 | — | |
| PIC-SepTraining Paradigm=In-Context Learning2026.04 | 6.3 | 7.2 | 7.9 | 8.2 | 8.6 | 7.6 | — | |
| DGCNNModel training/learning paradigm=Multi-task models: share backbone + multi-task heads, Venues=TOG' 192024.11 | 6.5 | 6.3 | 6.5 | 6.4 | 7.1 | 6.5 | — | |
| DGCNN†Training Strategy=Multitask†2024.04 | 6.5 | 6.3 | 6.5 | 6.4 | 7.1 | 6.5 | — | |
| DGCNNTraining Paradigm=Multitask, Pre-trained=No2026.04 | 6.5 | 6.3 | 6.5 | 6.4 | 7.1 | 6.5 | — | |
| PIC-SepModel training/learning paradigm=In-context learning models, Venues=NeurIPS'232024.11 | 6.9 | 6.3 | 7.2 | 7.9 | 8.2 | 7.3 | — | |
| ACTModel training/learning paradigm=Multi-task models: share backbone + multi-task heads, Venues=ICLR'232024.11 | 7.3 | 6.8 | 7 | 6.8 | 7.2 | 7 | — | |
| ACT‡Training Strategy=Multitask‡2024.04 | 7.3 | 6.8 | 7 | 6.8 | 7.2 | 7 | — | |
| ACTTraining Paradigm=Multitask, Pre-trained=Yes2026.04 | 7.3 | 6.8 | 7 | 6.8 | 7.2 | 7 | — | |
| PCP-MAETraining Paradigm=Multitask, Pre-trained=Yes2026.04 | 7.6 | 8.4 | 9.1 | 9.8 | 10.7 | 9.1 | — | |
| DGCNN‡Training Strategy=Multitask‡2024.04 | 8.2 | 8.3 | 8.4 | 8.8 | 9.2 | 8.6 | — | |
| PIC-S-SepModel training/learning paradigm=In-context learning models, Venues=Arxiv'242024.11 | 9.4 | 11.7 | 12.5 | 13.1 | 13.4 | 12 | — | |
| PIC-S-SepTraining Paradigm=In-Context Learning2026.04 | 9.4 | 11.7 | 12.5 | 13.1 | 13.4 | 12 | — | |
| UniPre3DTraining Paradigm=Multitask, Pre-trained=Yes2026.04 | 10.3 | 10.9 | 11.6 | 12.2 | 13.3 | 11.6 | — | |
| PCTModel training/learning paradigm=Multi-task models: share backbone + multi-task heads, Venues=CVM'212024.11 | 10.6 | 11.2 | 10.3 | 10.7 | 10.2 | 10.5 | — | |
| PCT†Training Strategy=Multitask†2024.04 | 11.2 | 10.3 | 10.7 | 10.2 | 10.5 | 10.6 | — | |
| PCTTraining Paradigm=Multitask, Pre-trained=No2026.04 | 11.2 | 10.3 | 10.7 | 10.2 | 10.5 | 10.6 | — | |
| MAMBA3D‡Training Strategy=Multitask‡2024.04 | 13.2 | 13.8 | 13.5 | 14.8 | 12.8 | 13.6 | — | |
| PCT‡Training Strategy=Multitask‡2024.04 | 14.5 | 12.2 | 12.4 | 12 | 11.8 | 12.6 | — | |
| PCMamba‡Training Strategy=Multitask‡2024.04 | 14.5 | 14.2 | 15.9 | 14.7 | 14.5 | 14.8 | — | |
| PointNetModel training/learning paradigm=Multi-task models: share backbone + multi-task heads, Venues=CVPR'172024.11 | 17.8 | 22 | 25.6 | 30.4 | 33.2 | 25.8 | — | |
| PointNet†Training Strategy=Multitask†2024.04 | 17.8 | 22 | 25.6 | 30.4 | 33.2 | 25.8 | — | |
| PointNetTraining Paradigm=Multitask, Pre-trained=No2026.04 | 17.8 | 22 | 25.6 | 30.4 | 33.2 | 25.8 | — | |
| ReConModel training/learning paradigm=Multi-task models: share backbone + multi-task heads, Venues=ICML 232024.11 | 20.4 | 24.5 | 27.2 | 29.2 | 32.5 | 26.9 | — | |
| ReCon‡Training Strategy=Multitask‡2024.04 | 20.4 | 24.5 | 27.2 | 29.2 | 32.5 | 26.9 | — | |
| ReConTraining Paradigm=Multitask, Pre-trained=Yes2026.04 | 20.4 | 24.5 | 27.2 | 29.2 | 32.5 | 26.9 | — | |
| I2P-MAEModel training/learning paradigm=Multi-task models: share backbone + multi-task heads, Venues=CVPR'232024.11 | 20.6 | 20.4 | 20.1 | 18.3 | 18.8 | 19.6 | — | |
| I2P-MAE‡Training Strategy=Multitask‡2024.04 | 20.6 | 20.4 | 20.1 | 18.3 | 18.8 | 19.6 | — | |
| I2P-MAETraining Paradigm=Multitask, Pre-trained=Yes2026.04 | 20.6 | 20.4 | 20.1 | 18.3 | 18.8 | 19.6 | — | |
| PointNet‡Training Strategy=Multitask‡2024.04 | 22.9 | 23.2 | 26.3 | 28.3 | 30 | 26.1 | — | |
| CopyModel training/learning paradigm=In-context learning models2024.11 | 149 | 155 | 157 | 155 | 155 | 154 | — | |
| CopyTraining Strategy=In-context Learning2024.04 | 149 | 155 | 157 | 155 | 155 | 154 | — | |
| Point-BERTModel training/learning paradigm=In-context learning models, Venues=CVPR 222024.11 | 292 | 293 | 298 | 296 | 299 | 296 | — | |
| Point-BERTTraining Strategy=In-context Learning2024.04 | 292 | 293 | 298 | 296 | 299 | 296 | — | |
| DeformPIC-8dFlops(G)=4.87, Param.(M)=29.842026.04 | — | — | — | — | — | — | 47.32 | |
| PIC-CatFlops(G)=11.31, Param.(M)=28.912026.04 | — | — | — | — | — | — | 52.89 | |
| PIC-SepFlops(G)=8.14, Param.(M)=28.912026.04 | — | — | — | — | — | — | 65.27 |