Research Article
Reconstructing 3D Model from Single-View Sketch with Deep Neural Network
Table 1
Segmentation accuracy (%) in terms of P-metric under different parameter settings on the SketchSeg dataset.
| Loss weight | Airplane | Bicycle | Candelabra | Chair | Four legs | Human | Lamp | Rifle | Table | Vase | Average |
| | 89.3 | 89.2 | 91.5 | 90.4 | 87.3 | 84.2 | 89.4 | 89.2 | 93.4 | 95.8 | 89.8 | | 92.1 | 91.6 | 93.1 | 91.8 | 87.6 | 84.3 | 89.6 | 90.4 | 88.7 | 92.1 | 90.1 | | 90.3 | 91.0 | 91.8 | 91.7 | 85.9 | 82.9 | 90.2 | 88.8 | 87.9 | 91.3 | 89.2 | | 88.7 | 89.6 | 89.4 | 85.3 | 83.6 | 80.6 | 86.9 | 83.5 | 81.6 | 88.5 | 85.8 | | 84.0 | 84.5 | 85.7 | 81.6 | 82.0 | 74.9 | 83.0 | 84.0 | 76.6 | 80.7 | 81.7 |
|
|