Scene Graph Generation On Visual Genome

평가 지표

Recall@100
Recall@20
Recall@50

평가 결과

이 벤치마크에서 각 모델의 성능 결과

모델 이름
Recall@100
Recall@20
Recall@50
Paper TitleRepository
LOGIN31.422.228.2Tackling the Challenges in Scene Graph Generation with Local-to-Global Interactions-
ADTrans--23.0Panoptic Scene Graph Generation with Semantics-Prototype Learning-
Graph-RCNN--11.4Graph R-CNN for Scene Graph Generation-
Causal-TDE--31.93Unbiased Scene Graph Generation from Biased Training-
NODIS-21.627.7NODIS: Neural Ordinary Differential Scene Understanding-
IETrans27.2-23.5Fine-Grained Scene Graph Generation with Data Transfer-
DLFE--25.4Recovering the Unbiased Scene Graphs from the Biased Ones-
ExpressiveSGG---Expressive Scene Graph Generation Using Commonsense Knowledge Infusion for Visual Understanding and Reasoning
SpeaQ (with reweighting)35.5-32.1Groupwise Query Specialization and Quality-Aware Multi-Assignment for Transformer-based Visual Relationship Detection-
VCTree--27.9Learning to Compose Dynamic Tree Structures for Visual Contexts-
GPS-Net--28.9GPS-Net: Graph Property Sensing Network for Scene Graph Generation-
KERN--27.1Knowledge-Embedded Routing Network for Scene Graph Generation-
SpeaQ (without reweighting)36.0-32.9Groupwise Query Specialization and Quality-Aware Multi-Assignment for Transformer-based Visual Relationship Detection-
C-bias---Biasing Like Human: A Cognitive Bias Framework for Scene Graph Generation-
CogTree---CogTree: Cognition Tree Loss for Unbiased Scene Graph Generation-
SG-EBM--31.74Energy-Based Learning for Scene Graph Generation-
MSDN--10.72Scene Graph Generation from Objects, Phrases and Region Captions-
KnowZRel---KnowZRel: Common Sense Knowledge-based Zero-Shot Relationship Retrieval for Generalised Scene Graph Generation
NeuSyRE---NeuSyRE: Neuro-Symbolic Visual Understanding and Reasoning Framework based on Scene Graph Enrichment
0 of 19 row(s) selected.
Scene Graph Generation On Visual Genome | SOTA | HyperAI초신경