Panoptic Scene Graph Generation On Psg
Métriques
R@20
mR@20
Résultats
Résultats de performance de divers modèles sur ce benchmark
Nom du modèle | R@20 | mR@20 | Paper Title | Repository |
|---|---|---|---|---|
| VCTree | 20.6 | 9.70 | Learning to Compose Dynamic Tree Structures for Visual Contexts | |
| ADTrans | 26.0 | 26.4 | Panoptic Scene Graph Generation with Semantics-Prototype Learning | |
| HiLo(R50) | 34.1 | 23.7 | HiLo: Exploiting High Low Frequency Relations for Unbiased Panoptic Scene Graph Generation | |
| MOTIFS | 20.0 | 9.10 | Neural Motifs: Scene Graph Parsing with Global Context | |
| PSGTR | 28.4 | 16.6 | Panoptic Scene Graph Generation | |
| IMP | 16.5 | 6.52 | Scene Graph Generation by Iterative Message Passing | |
| HiLo(SwinL) | 40.6 | 29.7 | HiLo: Exploiting High Low Frequency Relations for Unbiased Panoptic Scene Graph Generation | |
| PSGFormer | 18.0 | 14.8 | Panoptic Scene Graph Generation | |
| VLPrompt (R50) | 39.4 | 34.7 | VLPrompt: Vision-Language Prompting for Panoptic Scene Graph Generation |
0 of 9 row(s) selected.