Table 5 Comparative performance of different neural network types in the proposed approach
From: 3D tooth identification for forensic dentistry using deep learning
Scenario\Accuracy | 16 Categories | 8 Categories | 4 Categories |
|---|---|---|---|
Features | Â | Â | Â |
GRU with only vertices | 94.29% | 96.09% | 98.63% |
GRU with only triangles center | 94.79% | 97.69% | 98.76% |
GRU[vertices + triangle centers] | 98.01% | 98.88% | 99.75% |