Face Verification On Ijb B
Métriques
TAR @ FAR=0.001
TAR @ FAR=0.01
TAR@FAR=0.0001
Résultats
Résultats de performance de divers modèles sur ce benchmark
Tableau comparatif
| Nom du modèle | TAR @ FAR=0.001 | TAR @ FAR=0.01 | TAR@FAR=0.0001 |
|---|---|---|---|
| unified-negative-pair-generation-toward-well | 96.6 | 97.7% | 95.04 |
| adaface-quality-adaptive-margin-for-face | - | - | 95.67 |
| ghostvlad-for-set-based-face-recognition | - | 96.4% | - |
| vggface2-a-dataset-for-recognising-faces | 90.8 | 95.6% | - |
| killing-two-birds-with-one-stone-efficient | - | - | 96.71 |
| unified-negative-pair-generation-toward-well | 96.5 | 97.36% | 94.99 |
| unified-negative-pair-generation-toward-well | 96.5 | 97.63% | 95.21 |
| adaface-quality-adaptive-margin-for-face | - | - | 96.03 |
| faceposenet-making-a-case-for-landmark-free | - | 96.5% | - |
| adaface-quality-adaptive-margin-for-face | - | - | 95.84 |
| cluster-and-aggregate-face-recognition-with | 96.91 | - | 95.53 |
| qmagface-simple-and-accurate-quality-aware | 96.48 | 97.72% | 94.7 |