@inproceedings{yao-etal-2026-scenelm,
title = "{S}cene{LM}: 3{D}-Aware Language Models for Editable 3{D} Scene Synthesis",
author = "Yao, Xingbo and
Chen, Xiaoyu and
Zhang, Doudou and
Sheng, Mingzhi and
Cao, Boyuan and
Chen, Ying-Cong and
Xiong, Hui",
editor = "Liakata, Maria and
Moreira, Viviane P. and
Zhang, Jiajun and
Jurgens, David",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {ACL} 2026",
month = jul,
year = "2026",
address = "San Diego, California, United States",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-acl.2116/",
pages = "42615--42624",
ISBN = "979-8-89176-395-1",
abstract = "Synthesizing an editable 3D scene from a single RGB image is central to content creation, embodied-agent data generation, and AR/VR, yet remains challenging to achieve both high-fidelity reconstruction and convenient interactive editing. Existing geometry-based pipelines produce high-quality 3D results but are typically hard to refine without rerunning the full process, while LLM-driven procedural systems enable interactive tool use but are mostly text-driven and lack precise metric 3D understanding from images. We present SceneLM, a language-model-based framework that grounds 3D scene synthesis in visual evidence by recovering an executable metric 3D layout directly from a single image. Given an RGB image (and camera intrinsics when available), SceneLM outputs a JSON-form layout specifying each object{'}s category, 3D center, size, and discretized yaw, and then deterministically executes this layout with a tool suite to instantiate, place, and edit objects for iterative refinement. To train metric layout recovery at scale, we curate five datasets covering diverse indoor, outdoor, and tabletop scenes and convert heterogeneous 3D annotations into a unified instruction-tuning format. To improve numerical stability and metric accuracy while preserving the text interface, we augment autoregressive JSON generation with a lightweight geometry prediction branch and dual supervision. Experiments show that SceneLM substantially improves single-image 3D layout estimation over strong open and proprietary MLLM baselines, and yields higher-quality end-to-end scene generation in geometric consistency, physical plausibility, semantic alignment, and realism."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="yao-etal-2026-scenelm">
<titleInfo>
<title>SceneLM: 3D-Aware Language Models for Editable 3D Scene Synthesis</title>
</titleInfo>
<name type="personal">
<namePart type="given">Xingbo</namePart>
<namePart type="family">Yao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xiaoyu</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Doudou</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mingzhi</namePart>
<namePart type="family">Sheng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Boyuan</namePart>
<namePart type="family">Cao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ying-Cong</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hui</namePart>
<namePart type="family">Xiong</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: ACL 2026</title>
</titleInfo>
<name type="personal">
<namePart type="given">Maria</namePart>
<namePart type="family">Liakata</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Viviane</namePart>
<namePart type="given">P</namePart>
<namePart type="family">Moreira</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jiajun</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">David</namePart>
<namePart type="family">Jurgens</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">San Diego, California, United States</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-395-1</identifier>
</relatedItem>
<abstract>Synthesizing an editable 3D scene from a single RGB image is central to content creation, embodied-agent data generation, and AR/VR, yet remains challenging to achieve both high-fidelity reconstruction and convenient interactive editing. Existing geometry-based pipelines produce high-quality 3D results but are typically hard to refine without rerunning the full process, while LLM-driven procedural systems enable interactive tool use but are mostly text-driven and lack precise metric 3D understanding from images. We present SceneLM, a language-model-based framework that grounds 3D scene synthesis in visual evidence by recovering an executable metric 3D layout directly from a single image. Given an RGB image (and camera intrinsics when available), SceneLM outputs a JSON-form layout specifying each object’s category, 3D center, size, and discretized yaw, and then deterministically executes this layout with a tool suite to instantiate, place, and edit objects for iterative refinement. To train metric layout recovery at scale, we curate five datasets covering diverse indoor, outdoor, and tabletop scenes and convert heterogeneous 3D annotations into a unified instruction-tuning format. To improve numerical stability and metric accuracy while preserving the text interface, we augment autoregressive JSON generation with a lightweight geometry prediction branch and dual supervision. Experiments show that SceneLM substantially improves single-image 3D layout estimation over strong open and proprietary MLLM baselines, and yields higher-quality end-to-end scene generation in geometric consistency, physical plausibility, semantic alignment, and realism.</abstract>
<identifier type="citekey">yao-etal-2026-scenelm</identifier>
<location>
<url>https://aclanthology.org/2026.findings-acl.2116/</url>
</location>
<part>
<date>2026-07</date>
<extent unit="page">
<start>42615</start>
<end>42624</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T SceneLM: 3D-Aware Language Models for Editable 3D Scene Synthesis
%A Yao, Xingbo
%A Chen, Xiaoyu
%A Zhang, Doudou
%A Sheng, Mingzhi
%A Cao, Boyuan
%A Chen, Ying-Cong
%A Xiong, Hui
%Y Liakata, Maria
%Y Moreira, Viviane P.
%Y Zhang, Jiajun
%Y Jurgens, David
%S Findings of the Association for Computational Linguistics: ACL 2026
%D 2026
%8 July
%I Association for Computational Linguistics
%C San Diego, California, United States
%@ 979-8-89176-395-1
%F yao-etal-2026-scenelm
%X Synthesizing an editable 3D scene from a single RGB image is central to content creation, embodied-agent data generation, and AR/VR, yet remains challenging to achieve both high-fidelity reconstruction and convenient interactive editing. Existing geometry-based pipelines produce high-quality 3D results but are typically hard to refine without rerunning the full process, while LLM-driven procedural systems enable interactive tool use but are mostly text-driven and lack precise metric 3D understanding from images. We present SceneLM, a language-model-based framework that grounds 3D scene synthesis in visual evidence by recovering an executable metric 3D layout directly from a single image. Given an RGB image (and camera intrinsics when available), SceneLM outputs a JSON-form layout specifying each object’s category, 3D center, size, and discretized yaw, and then deterministically executes this layout with a tool suite to instantiate, place, and edit objects for iterative refinement. To train metric layout recovery at scale, we curate five datasets covering diverse indoor, outdoor, and tabletop scenes and convert heterogeneous 3D annotations into a unified instruction-tuning format. To improve numerical stability and metric accuracy while preserving the text interface, we augment autoregressive JSON generation with a lightweight geometry prediction branch and dual supervision. Experiments show that SceneLM substantially improves single-image 3D layout estimation over strong open and proprietary MLLM baselines, and yields higher-quality end-to-end scene generation in geometric consistency, physical plausibility, semantic alignment, and realism.
%U https://aclanthology.org/2026.findings-acl.2116/
%P 42615-42624
Markdown (Informal)
[SceneLM: 3D-Aware Language Models for Editable 3D Scene Synthesis](https://aclanthology.org/2026.findings-acl.2116/) (Yao et al., Findings 2026)
ACL
- Xingbo Yao, Xiaoyu Chen, Doudou Zhang, Mingzhi Sheng, Boyuan Cao, Ying-Cong Chen, and Hui Xiong. 2026. SceneLM: 3D-Aware Language Models for Editable 3D Scene Synthesis. In Findings of the Association for Computational Linguistics: ACL 2026, pages 42615–42624, San Diego, California, United States. Association for Computational Linguistics.