Single-image 3D shape reconstruction on ShapeNet Core v1 (test)
74.84F1 (tau)Mesh R-CNN (Ours Best)
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Mesh R-CNN (Ours Best)variant=Best2019.06 | 74.84 | 0.306 | 85.75 | |
| Mesh R-CNN (Ours Pretty)variant=Pretty2019.06 | 69.83 | 0.391 | 81.76 | |
| Pixel2Meshstatus=model released by authors2019.06 | 67.89 | 0.463 | 79.88 | |
| GEOMetrics2019.06 | 67.37 | — | — | |
| MVD2019.06 | 66.39 | — | — | |
| Pixel2Meshstatus=reported in paper2019.06 | 59.72 | 0.591 | 74.19 | |
| PSG2019.06 | 48.58 | 0.593 | 69.78 | |
| 3D-R2N22019.06 | 39.01 | 1.445 | 54.62 | |
| N3MR2019.06 | 33.8 | 2.629 | 47.72 |