Camera Pose Estimation on ScanNet-50 (700 Images)
0.08ATEpi^3
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| pi^3Dense settings=true, Full-attention baseline=true, Spd.=1.0×, Backbone=pi^32026.07 | 0.08 | 2.15 | 0.01 | 0.37 | |
| Fast pi^3Dense settings=true, Full-attention baseline=false, Spd.=2.3×, Backbone=pi^32026.07 | 0.09 | 2.39 | 0.02 | 0.45 | |
| A pi^3Dense settings=true, Full-attention baseline=false, Spd.=6.7×, Backbone=pi^32026.07 | 0.09 | 2.37 | 0.02 | 0.4 | |
| SAF3RDense settings=true, Full-attention baseline=false, Spd.=7.3×, Backbone=pi^32026.07 | 0.09 | 2.44 | 0.01 | 0.39 | |
| FastVGGTDense settings=true, Full-attention baseline=false, Spd.=3.4×, Backbone=VGGT2026.07 | 0.14 | 3.72 | 0.03 | 0.73 | |
| Sparse pi^3Dense settings=true, Full-attention baseline=false, Spd.=3.1×, Backbone=pi^32026.07 | 0.14 | 4.98 | 0.02 | 0.43 | |
| SAF3RDense settings=true, Full-attention baseline=false, Spd.=4.7×, Backbone=VGGT2026.07 | 0.17 | 4.47 | 0.04 | 1.03 | |
| Co-MeDense settings=true, Full-attention baseline=false, Spd.=4.0×, Backbone=VGGT2026.07 | 0.18 | 4.58 | 0.06 | 1.2 | |
| VGGTDense settings=true, Full-attention baseline=true, Spd.=1.0×, Backbone=VGGT2026.07 | 0.19 | 4.63 | 0.04 | 0.97 | |
| AVGGTDense settings=true, Full-attention baseline=false, Spd.=4.6×, Backbone=VGGT2026.07 | 0.2 | 5.37 | 0.05 | 1.6 | |
| SparseVGGTDense settings=true, Full-attention baseline=false, Spd.=2.8×, Backbone=VGGT2026.07 | 0.21 | 5.85 | 0.06 | 1.65 |