Table 4 Our dental classification results compared to state-of-the-art methods
From: 3D tooth identification for forensic dentistry using deep learning
Scenario\Accuracy | 16 Categories | 8 Categories | 4 Categories |
|---|---|---|---|
Network type | Â | Â | Â |
CNN | 76.45% | 81.41% | 90.33% |
Vanilla RNN | 90.08% | 93.30% | 98.02% |
LSTM | 93.92% | 96.28% | 98.14% |
GRU | 98.01% | 98.88% | 99.75% |