Table 4 The proposed framework results compared to the recent state-of-the-art approaches on the ShapeNet dataset.
From: 3D reconstruction from 2D multi-view dental 2D images based on EfficientNetB0 model
Year | Research | Network | IOU (%) | F1-score (%) | CD (%) | EMD (%) |
|---|---|---|---|---|---|---|
2016 | Choy et al.42 | 3D-R2N2 | 63.4 | – | – | – |
2021 | Yang et al.57 | LSTM shape encoder | 38.7 | 14.3 | – | – |
2024 | Kalitsios et al.55 | Multi-manifold attention | 78.12 | 56 | – | – |
2025 | Liu et al.56 | Prior-guided adaptive probabilistic network for single-view 3D reconstruction | – | 54.42 | 2.62 | 5.99 |
2025 | Proposed methodology | EfficientNetB0 | 89.98 | 94.11 | 0.41 | 0.24 |
