SceneReVis: A Self-Reflective Vision-Grounded Framework for 3D Indoor Scene Synthesis via Multi-Turn RL

arXiv:2602.09432 · cs.CV · Submitted 2026-02-10 · Read on arXiv

cs.CV

Submitted: 2026-02-10

Updated: 2026-10-15

Comments: 9 pages; accepted to ACM Multimedia 2026 (Oral). Updated to the camera-ready version

DOI: 10.1145/3767308.3835940

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Terminology

• Fidelity — 1× in this paper · explained in 2 episodes
• Fine-Tuning — 1× in this paper · explained in 5 episodes
• Generalization — 1× in this paper · explained in 13 episodes
• Hallucination — 1× in this paper · explained in 4 episodes
• Reinforcement Learning — 1× in this paper · explained in 4 episodes
• Two-stage training — 1× in this paper · explained in 2 episodes

Sources

• Qwen3-VL Technical Report
• Qwen2.5-VL Technical Report
• ReSpace: Text-Driven Autoregressive 3D Indoor Scene Synthesis and Editing
• SceneNAT: Masked Generative Modeling for Language-Guided Indoor Scene Synthesis
• Agentic Reinforced Policy Optimization
• ReTool: Reinforcement Learning for Strategic Tool Use in LLMs
• DisCo-Layout: Disentangling and Coordinating Semantic and Physical Refinement in a Multi-Agent Framework for 3D Indoor Layout Synthesis
• DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning
• InstructScene: Instruction-Driven 3D Indoor Scene Synthesis with Semantic Graph Prior
• Scenethesis: A Language and Vision Agentic Framework for 3D Scene Generation
• MetaSpatial: Reinforcing 3D Spatial Reasoning in VLMs for the Metaverse
• Direct Numerical Layout Generation for 3D Indoor Scene Synthesis via Spatial Reasoning
• Agentic Reasoning and Tool Integration for LLMs via Reinforcement Learning
• A Practitioner's Guide to Multi-turn Agentic Reinforcement Learning

Related papers

• Loss Knows Best: Detecting Annotation Errors in Videos via Loss Trajectories
• AnchorWeave: World-Consistent Video Generation with Retrieved Local Spatial Memories
• Benchmarking the Robustness of Foundation Models for Mammography under Domain Shift
• MambaX-Net: Dual-Input Mamba-Enhanced Cross-Attention Network for Longitudinal MRI Segmentation
• TeleOCR: Navigating Document Parsing Across Digital and Camera-Captured Documents
• A Survey on Efficient Vision-Language-Action Models