Scene Graph Generation

A scene graph is a structured representation of an image, where nodes in a scene graph correspond to object bounding boxes with their object categories, and edges correspond to their pairwise relationships between objects. The task of Scene Graph Generation is to generate a visually-grounded scene graph that most accurately correlates with an image.

Source: Scene Graph Generation by Iterative Message Passing

Papers

Recently Added Most Hyped Most Active Needs Verification Most Verified

Showing 1–10 of 318 papers

Title	Date	Tasks	Status	Hype
4D Panoptic Scene Graph Generation	May 16, 2024	4D Panoptic SegmentationGraph Generation	CodeCode Available	3
REACT: Real-time Efficiency and Accuracy Compromise for Tradeoffs in Scene Graph Generation	May 25, 2024	Graph GenerationObject	CodeCode Available	2
HiKER-SGG: Hierarchical Knowledge Enhanced Robust Scene Graph Generation	Mar 18, 2024	Scene Graph Generation	CodeCode Available	2
MM-OR: A Large Multimodal Operating Room Dataset for Semantic Understanding of High-Intensity Surgical Environments	Mar 4, 2025	2D Panoptic SegmentationGraph Generation	CodeCode Available	2
Open World Scene Graph Generation using Vision Language Models	Jun 9, 2025	Graph GenerationScene Graph Generation	CodeCode Available	2
Panoptic Scene Graph Generation	Jul 22, 2022	BenchmarkingPanoptic Scene Graph Generation	CodeCode Available	2
EGTR: Extracting Graph from Transformer for Scene Graph Generation	Apr 2, 2024	Graph GenerationMulti-Task Learning	CodeCode Available	2
From Pixels to Graphs: Open-Vocabulary Scene Graph Generation with Vision-Language Models	Apr 1, 2024	Graph GenerationImage to text	CodeCode Available	2
Learning to Compose Dynamic Tree Structures for Visual Contexts	Dec 5, 2018	Graph GenerationPanoptic Scene Graph Generation	CodeCode Available	2
RelationField: Relate Anything in Radiance Fields	Dec 18, 2024	3d scene graph generationGraph Generation	CodeCode Available	2

Show:10 25 50

← PrevPage 1 of 32Next →

All datasets Visual Genome 4D-OR 3R-Scan VRD GQA MM-OR MS-COCO

Benchmark Results

#	Model	Metric	Claimed	Verified	Status
1	ExpressiveSGG	R@100	39.12	—	Unverified
2	NeuSyRE	R@100	39.1	—	Unverified
3	KnowZRel	zR@100	35.65	—	Unverified
4	SpeaQ (without reweighting)	Recall@50	32.9	—	Unverified
5	SpeaQ (with reweighting)	Recall@50	32.1	—	Unverified
6	Causal-TDE	Recall@50	31.93	—	Unverified
7	SG-EBM	Recall@50	31.74	—	Unverified
8	GPS-Net	Recall@50	28.9	—	Unverified
9	LOGIN	Recall@50	28.2	—	Unverified
10	VCTree	Recall@50	27.9	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1	ORacle	F1	0.91	—	Unverified
2	MM2SG	F1	0.9	—	Unverified
3	Pix2SG	F1	0.9	—	Unverified
4	LABRAD-OR	F1	0.88	—	Unverified
5	4D-OR baseline	F1	0.75	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1	SceneGraphFusion	Top-5 Accuracy	0.87	—	Unverified
2	3DSSG [Wald2020_3dssg]	Top-5 Accuracy	0.66	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1	FactorizableNet	Recall@50	18.32	—	Unverified
2	VRD	Recall@50	18.16	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1	KnowZRel	zR@100	29.56	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1	MM2SG	Macro F1	0.53	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1	NeuSyRE	R@100	38.5	—	Unverified