Zero Shot Skeletal Action Recognition On Ntu 1
Métriques
Accuracy (10 unseen classes)
Accuracy (24 unseen classes)
Résultats
Résultats de performance de divers modèles sur ce benchmark
Tableau comparatif
| Nom du modèle | Accuracy (10 unseen classes) | Accuracy (24 unseen classes) |
|---|---|---|
| part-aware-unified-representation-of-language-1 | 71.95 | 52.01 |
| fine-grained-side-information-guided-dual | 63.30 | 44.30 |
| tdsm-triplet-diffusion-for-skeleton-text | 74.15 | 65.06 |
| zero-shot-skeleton-based-action-recognition | 65.74 | 45.30 |
| sa-dvae-improving-zero-shot-skeleton-based | 68.77 | 46.12 |
| syntactically-guided-generative-embeddings-1 | 62.69 | 38.70 |
| generalized-zero-and-few-shot-learning-via | 59.53 | 35.77 |
| sa-dvae-improving-zero-shot-skeleton-based | - | - |