A learnable 3D scene graph representation that injects semantic relation embeddings between objects improves LLM performance on 3D grounding, captioning, and question answering.
Scanqa: 3d question answering for spatial scene understanding
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CV 1years
2024 1verdicts
CONDITIONAL 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
3DGraphLLM: Combining Semantic Graphs and Large Language Models for 3D Scene Understanding
A learnable 3D scene graph representation that injects semantic relation embeddings between objects improves LLM performance on 3D grounding, captioning, and question answering.