Facial Landmark Detection On 300W
Métriques
NME
Résultats
Résultats de performance de divers modèles sur ce benchmark
Tableau comparatif
| Nom du modèle | NME |
|---|---|
| deep-structured-prediction-for-facial-1 | 3.30 |
| face-alignment-across-large-poses-a-3d | 5.76 |
| cascaded-dual-vision-transformer-for-accurate | 2.85 |
| face-alignment-using-a-3d-deeply-initialized | 3.13 |
| style-aggregated-network-for-facial-landmark | 3.98 |
| faceposenet-making-a-case-for-landmark-free | - |
| pose-invariant-face-alignment-with-a-single | 6.30 |
| anchorface-an-anchor-based-facial-landmark | 3.12 |
| cascade-of-encoder-decoder-cnns-with-learned | 3.3 |
| adaloss-adaptive-loss-function-for-landmark | 3.31 |
| shape-preserving-facial-landmarks-with-graph | 2.99 |
| a-deeply-initialized-coarse-to-fine-ensemble | 3.24 |
| fiducial-focus-augmentation-for-facial | 2.89 |
| face-alignment-across-large-poses-a-3d | 7.01 |
| teacher-supervises-students-how-to-learn-from | 3.49 |