Registration on ShapeNet In-Context
3.4L1 CD Error (x1000)PIC-Sep + MICAS
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| PIC-Sep + MICASModel training/learning paradigm=In-context learning models, Venues=Ours2024.11 | 3.4 | 3.6 | 3.7 | 3.8 | 4 | 3.7 | |
| ACTModel training/learning paradigm=Task-specific models (trained separately), Venues=ICLR'232024.11 | 5.1 | 5.6 | 5.9 | 6 | 7 | 5.9 | |
| ACTTraining Strategy=Task-Specific2024.04 | 5.1 | 5.6 | 5.9 | 6 | 7 | 5.9 | |
| PointNetModel training/learning paradigm=Task-specific models (trained separately), Venues=CVPR'172024.11 | 5.3 | 5.9 | 6.9 | 7.7 | 8.5 | 6.9 | |
| PCTModel training/learning paradigm=Task-specific models (trained separately), Venues=CVM'212024.11 | 5.3 | 5.7 | 6.3 | 6.9 | 7.2 | 6.3 | |
| PointNetTraining Strategy=Task-Specific2024.04 | 5.3 | 5.9 | 6.9 | 7.7 | 8.5 | 6.9 | |
| PCTTraining Strategy=Task-Specific2024.04 | 5.3 | 5.7 | 6.3 | 6.9 | 7.2 | 6.3 | |
| PIC-Cat + MICASModel training/learning paradigm=In-context learning models, Venues=Ours2024.11 | 5.7 | 6.5 | 9.1 | 12.5 | 15.4 | 9.8 | |
| PIC-S-SepModel training/learning paradigm=In-context learning models, Venues=Arxiv'242024.11 | 6 | 6.1 | 7.6 | 6.7 | 7.3 | 6.7 | |
| DGCNNModel training/learning paradigm=Task-specific models (trained separately), Venues=TOG' 192024.11 | 6.2 | 6.7 | 7.3 | 7.4 | 7.7 | 7.1 | |
| DGCNNTraining Strategy=Task-Specific2024.04 | 6.2 | 6.7 | 7.3 | 7.4 | 7.7 | 7.1 | |
| PIC-SepModel training/learning paradigm=In-context learning models, Venues=NeurIPS'232024.11 | 8.6 | 9.2 | 10.2 | 11.3 | 12.4 | 10.3 | |
| PIC++Training Strategy=In-context Learning, Aggregation=-Cat2024.04 | 9.7 | 11.6 | 12.8 | 15.9 | 18 | 13.6 | |
| PIC-CatModel training/learning paradigm=In-context learning models, Venues=NeurIPS'232024.11 | 10 | 11.4 | 13.8 | 16.9 | 18.6 | 14.1 | |
| PICTraining Strategy=In-context Learning, Aggregation=-Cat2024.04 | 10 | 11.4 | 13.8 | 16.9 | 18.6 | 14.1 | |
| Point-MAEModel training/learning paradigm=Multi-task models: share backbone + multi-task heads, Venues=ECCV'222024.11 | 11.4 | 12.8 | 14.8 | 16 | 16.9 | 14.5 | |
| PointMAE‡Training Strategy=Multitask‡2024.04 | 11.4 | 12.8 | 14.8 | 16 | 16.9 | 14.5 | |
| ACTModel training/learning paradigm=Multi-task models: share backbone + multi-task heads, Venues=ICLR'232024.11 | 12.2 | 14.4 | 19.4 | 25.5 | 29 | 20.1 | |
| DGCNNModel training/learning paradigm=Multi-task models: share backbone + multi-task heads, Venues=TOG' 192024.11 | 12.5 | 14.9 | 17.9 | 19.7 | 20.7 | 17.1 | |
| DGCNN†Training Strategy=Multitask†2024.04 | 12.5 | 14.9 | 17.9 | 19.7 | 20.7 | 17.1 | |
| PIC-S-CatModel training/learning paradigm=In-context learning models, Venues=Arxiv'242024.11 | 12.8 | 15.8 | 23.9 | 31.2 | 36.9 | 24.1 | |
| ReConModel training/learning paradigm=Multi-task models: share backbone + multi-task heads, Venues=ICML 232024.11 | 14.7 | 16.3 | 19.2 | 21.5 | 22.5 | 18.8 | |
| PCTModel training/learning paradigm=Multi-task models: share backbone + multi-task heads, Venues=CVM'212024.11 | 24.4 | 26 | 29.6 | 32.8 | 34.7 | 29.5 | |
| PCT†Training Strategy=Multitask†2024.04 | 24.4 | 26 | 29.6 | 32.8 | 34.7 | 29.5 | |
| PointNetModel training/learning paradigm=Multi-task models: share backbone + multi-task heads, Venues=CVPR'172024.11 | 25.4 | 22.6 | 24.9 | 25.7 | 26.9 | 25.1 | |
| PointNet†Training Strategy=Multitask†2024.04 | 25.4 | 22.6 | 24.9 | 25.7 | 26.9 | 25.1 | |
| I2P-MAEModel training/learning paradigm=Multi-task models: share backbone + multi-task heads, Venues=CVPR'232024.11 | 32.5 | 31.3 | 31.1 | 31.6 | 31.2 | 31.5 | |
| CopyModel training/learning paradigm=In-context learning models2024.11 | 155 | 157 | 156 | 148 | 154 | 154 | |
| Point-BERTModel training/learning paradigm=In-context learning models, Venues=CVPR 222024.11 | 291 | 295 | 294 | 295 | 298 | 294 | |
| DeformPICTraining Paradigm=In-Context Learning2026.04 | 1,900 | 1,900 | 1,900 | 2,000 | 2,100 | 2,000 | |
| ACTTraining Paradigm=Task-Specific2026.04 | 5,100 | 5,600 | 5,900 | 6,000 | 7,000 | 5,900 | |
| PointNetTraining Paradigm=Task-Specific2026.04 | 5,300 | 5,900 | 6,900 | 7,700 | 8,500 | 6,900 | |
| PCTTraining Paradigm=Task-Specific2026.04 | 5,300 | 5,700 | 6,300 | 6,900 | 7,200 | 6,300 | |
| PIC-S-SepTraining Paradigm=In-Context Learning2026.04 | 6,000 | 6,100 | 7,600 | 6,700 | 7,300 | 6,700 | |
| DGCNNTraining Paradigm=Task-Specific2026.04 | 6,200 | 6,700 | 7,300 | 7,400 | 7,700 | 7,100 | |
| PIC-SepTraining Paradigm=In-Context Learning2026.04 | 8,600 | 9,200 | 10,200 | 11,300 | 12,400 | 10,300 | |
| PIC-CatTraining Paradigm=In-Context Learning2026.04 | 10,000 | 11,400 | 13,800 | 16,900 | 18,600 | 14,100 | |
| Point-MAETraining Paradigm=Multitask, Pre-trained=Yes2026.04 | 11,400 | 12,800 | 14,800 | 16,000 | 16,900 | 14,500 | |
| ACTTraining Paradigm=Multitask, Pre-trained=Yes2026.04 | 12,200 | 14,400 | 19,400 | 25,500 | 29,000 | 20,100 | |
| DGCNNTraining Paradigm=Multitask, Pre-trained=No2026.04 | 12,500 | 14,900 | 17,900 | 19,700 | 20,700 | 17,100 | |
| PIC-S-CatTraining Paradigm=In-Context Learning2026.04 | 12,800 | 15,800 | 23,900 | 31,200 | 36,900 | 24,100 | |
| ReConTraining Paradigm=Multitask, Pre-trained=Yes2026.04 | 14,700 | 16,300 | 19,200 | 21,500 | 22,500 | 18,800 | |
| PCP-MAETraining Paradigm=Multitask, Pre-trained=Yes2026.04 | 17,300 | 25,600 | 36,300 | 45,100 | 50,200 | 34,900 | |
| UniPre3DTraining Paradigm=Multitask, Pre-trained=Yes2026.04 | 18,200 | 26,300 | 36,500 | 44,900 | 49,700 | 35,100 | |
| PCTTraining Paradigm=Multitask, Pre-trained=No2026.04 | 24,400 | 26,000 | 29,600 | 32,800 | 34,700 | 29,500 | |
| PointNetTraining Paradigm=Multitask, Pre-trained=No2026.04 | 25,400 | 22,600 | 24,900 | 25,700 | 26,900 | 25,100 | |
| I2P-MAETraining Paradigm=Multitask, Pre-trained=Yes2026.04 | 32,500 | 31,300 | 31,100 | 31,600 | 31,200 | 31,500 |