Single-View Reconstruction on ShapeNet
68.4plaPix2Vox-A
Evaluation results
| Method | Links | |||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Pix2Vox-A3D CAD models used=44k2019.01 | 68.4 | 61.6 | 79.2 | 85.4 | 56.7 | 53.7 | 44.3 | 71.4 | — | 70.9 | 60.1 | 77.6 | 59.4 | 66.1 | 61.5 | |
| PSGNadditional_input=object masks, 3D CAD models used=220k2019.01 | 60.1 | 55 | 77.1 | 83.1 | 54.4 | 55.2 | 46.2 | 73.7 | — | 70.8 | 60.6 | 74.9 | 61.1 | 64 | 60.4 | |
| Pix2Vox-F3D CAD models used=44k2019.01 | 60 | 53.8 | 76.5 | 83.7 | 53.5 | 51.1 | 43.5 | 70.7 | — | 68.7 | 58.7 | 77 | 58.2 | 63.4 | 59.8 | |
| OGN3D CAD models used=44k2019.01 | 58.7 | 48.1 | 72.9 | 82.8 | 48.3 | 50.2 | 39.8 | 63.7 | — | 64.6 | 53.6 | 70.2 | 63.2 | 59.6 | 59.3 | |
| DRCtraining_testing_mode=per category, 3D CAD models used=44k2019.01 | 57.1 | 45.3 | 63.5 | 75.5 | 46.9 | 41.9 | 41.5 | 60.9 | — | 60.6 | 42.4 | 41.3 | 55.6 | 54.5 | 60.8 | |
| Pixel2mesh2022.02 | 53.6 | 51.4 | 48.5 | 46.9 | 57.7 | 52.8 | 68.7 | 61.7 | — | 53.4 | 51.3 | 42.2 | 54.8 | 52.7 | 42.1 | |
| 3D-R2N23D CAD models used=44k2019.01 | 51.3 | 42.1 | 71.6 | 79.8 | 46.6 | 46.8 | 38.1 | 66.2 | — | 62.8 | 51.3 | 66.1 | 51.3 | 56 | 54.4 | |
| 3DR2N22022.02 | 49.4 | 48 | 42.5 | 47.3 | 57.5 | 58.5 | 106.4 | 59.6 | — | 47.2 | 52.9 | 43.7 | 50.7 | 54.1 | 40.2 | |
| Pix2Vox++2022.02 | 36.5 | 44 | 39.9 | 34.8 | 39.7 | 44 | 56.3 | 48.4 | — | 41.2 | 40.1 | 36.8 | 42.8 | 41.7 | 37.8 | |
| Pix2Vox2022.02 | 34.8 | 44.7 | 43.9 | 35.6 | 40.4 | 44.7 | 56.6 | 51 | — | 43.7 | 42.9 | 38.4 | 41.4 | 42.8 | 38 | |
| OccNet2022.02 | 31.9 | 33.1 | 35.4 | 36.9 | 40.8 | 48.4 | 75.5 | 54.7 | — | 39.7 | 37.4 | 31.6 | 44.3 | 41.5 | 29.7 | |
| PSGN2022.02 | 27.8 | 37.3 | 41.2 | 32.7 | 46.8 | 47.4 | 56 | 56.2 | — | 44.4 | 38.1 | 38.1 | 38.4 | 40.7 | 25.3 | |
| DPM2022.02 | 26.4 | 35.6 | 34.6 | 32.3 | 41.5 | 43.5 | 51.8 | 51.4 | — | 41.5 | 37.1 | 34.2 | 35.2 | 37.6 | 24.1 | |
| AtlasNet2022.02 | 26 | 32 | 36.6 | 30.7 | 40.9 | 41.6 | 49.8 | 49.1 | — | 38 | 33.6 | 32 | 34 | 35.9 | 22 | |
| 3DAttriFlow2022.02 | 21.1 | 27.1 | 26.6 | 25 | 33.3 | 36 | 45.5 | 41.6 | — | 32.4 | 28.5 | 26.6 | 29.6 | 30.2 | 19.4 | |
| SnowflakeNet (Ours)2022.02 | 19.9 | 25.4 | 25.2 | 24.4 | 31.3 | 33.7 | 43.4 | 39.8 | — | 30.9 | 27.1 | 24.5 | 28 | 28.6 | 18.4 | |
| PSGMetric=Chamfer Distance, Number of points=1024, Post-processing=ICP alignment, Scaling factor=10^32018.02 | 3.36 | 4.31 | 8.51 | 8.63 | 6.35 | 6.47 | 7.66 | 15.9 | 1.58 | 6.92 | 3.93 | 3.76 | 5.94 | 6.41 | — | |
| Bansal et al.Metric=Chamfer Distance, Number of points=1024, Post-processing=ICP alignment, Scaling factor=10^32018.02 | 2.91 | 4.39 | 6.01 | 4.45 | 7.24 | 5.95 | 7.42 | 10.4 | 1.83 | 6.65 | 4.83 | 4.66 | 4.65 | 5.5 | — | |
| AtlasNetMetric=Chamfer Distance, Number of points=1024, Post-processing=ICP alignment, Scaling factor=10^32018.02 | 2.54 | 3.91 | 5.39 | 4.18 | 6.77 | 6.71 | 7.24 | 8.18 | 1.63 | 6.76 | 4.35 | 3.91 | 4.91 | 5.11 | — | |
| AtlasNetMetric=Metro distance, Number of points=1024, Post-processing=ICP alignment, Scaling factor=102018.02 | 1.31 | 1.89 | 1.8 | 2.04 | 2.11 | 1.68 | 2.81 | 2.39 | 1.57 | 1.78 | 2.28 | 1.03 | 1.84 | 1.89 | — |