OptiScene：通过规模化人工对齐数据合成与多阶段偏好优化的LLM驱动室内场景布局生成
OptiScene: LLM-driven Indoor Scene Layout Generation via Scaled Human-aligned Data Synthesis and Multi-Stage Preference Optimization
AI总结：
本文提出3D-SynthPlace大规模合成布局数据集和开源LLM OptiScene，通过监督微调与多轮直接偏好优化两阶段训练，显著提升室内布局生成质量并优于现有基线方法。
摘要：
自动室内布局生成因其在室内设计、虚拟环境构建和具身人工智能中的潜力而受到越来越多的关注。现有方法分为两类：利用专有LLM服务（如GPT API）的提示驱动方法，以及基于扩散模型在布局数据上训练的学习方法。提示驱动方法通常存在空间不一致性和高计算成本的问题，而学习方法通常受限于粗糙的关系图和有限的数据集，限制了它们对不同房间类别的泛化能力。在本文中，我们重新审视了基于LLM的室内布局生成，并提出了3D-SynthPlace，这是一个大规模数据集，结合了通过“GPT合成、人工检查”流程生成的合成布局，并从3D-Front数据集升级而来。3D-SynthPlace包含近17,000个场景，覆盖四种常见房间类型——卧室、客厅、厨房和浴室——并丰富了多样化的物体和高级空间标注。我们进一步提出了OptiScene，这是一个针对室内布局生成优化的强大开源LLM，基于我们的3D-SynthPlace数据集通过两阶段训练进行微调。对于预热阶段I，我们采用监督微调（SFT），教模型首先生成高级空间描述，然后条件性地预测具体物体摆放。对于强化阶段II，为了更好地使生成的布局与人类设计偏好对齐，我们应用多轮直接偏好优化（DPO），显著提高了布局质量和生成成功率。大量实验表明，OptiScene优于传统的提示驱动和学习基线方法。此外，OptiScene在场景编辑和机器人导航等交互式任务中展现出有前景的潜力。
英文摘要：
Automatic indoor layout generation has attracted increasing attention due to its potential in interior design, virtual environment construction, and embodied AI. Existing methods fall into two categories: prompt-driven approaches that leverage proprietary LLM services (e.g., GPT APIs) and learning-based methods trained on layout data upon diffusion-based models. Prompt-driven methods often suffer from spatial inconsistency and high computational costs, while learning-based methods are typically constrained by coarse relational graphs and limited datasets, restricting their generalization to diverse room categories. In this paper, we revisit LLM-based indoor layout generation and present 3D-SynthPlace, a large-scale dataset that combines synthetic layouts generated via a 'GPT synthesize, Human inspect' pipeline, upgraded from the 3D-Front dataset. 3D-SynthPlace contains nearly 17,000 scenes, covering four common room types -- bedroom, living room, kitchen, and bathroom -- enriched with diverse objects and high-level spatial annotations. We further introduce OptiScene, a strong open-source LLM optimized for indoor layout generation, fine-tuned based on our 3D-SynthPlace dataset through our two-stage training. For the warum-up stage I, we adopt supervised fine-tuning (SFT), which is taught to first generate high-level spatial descriptions then conditionally predict concrete object placements. For the reinforcing stage II, to better align the generated layouts with human design preferences, we apply multi-turn direct preference optimization (DPO), which significantly improving layout quality and generation success rates. Extensive experiments demonstrate that OptiScene outperforms traditional prompt-driven and learning-based baselines. Moreover, OptiScene shows promising potential in interactive tasks such as scene editing and robot navigation.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
