提出STAR框架，解决3D场景理解中跨模态拓扑差异带来的专家分配难题。
STAR: A Spatial-Topology Aware Routing Framework for Generalizable 3D Scene Understanding
- 引入多属性自监督预训练，锚定跨域结构先验
- 在ScanNet上达80.1% mIoU，S3DIS上77.2% mIoU
- 适合需要泛化3D场景理解的多模态系统研究者
构建统一的3D场景理解模型长期受限于不同传感器模态间的拓扑差异。尽管混合专家（MoE）架构适用于多领域3D理解，但传统仅依赖特征的MoE路由器在语义监督下可能忽略局部采样拓扑，导致语义一致性与几何异质性共存时专家分配困难。为此，我们提出空间-拓扑感知路由框架STAR。具体地，设计包含拓扑与纹理变化的多属性自监督预训练分支，以锚定跨域结构先验。在此基础上，构建领域感知专家分支，包含两种机制：基于领域-空间引导的路由（DSR），从空间上下文中捕捉局部拓扑变化；熵控制的动态分配（EDA），根据路由不确定性调整激活专家数量。两者结合实现稳定跨域表征学习与自适应专家分配。大量实验覆盖室内与室外场景，验证了STAR有效性。在ScanNet验证集上达80.1% mIoU，S3DIS上达77.2%，持续优于强基线。代码已公开于项目页面（https://xmw666.github.io/STAR/）。
原文摘要 · Abstract (English)
Constructing a unified 3D scene understanding model has long been hindered by the topological discrepancies across sensor modalities. While applying the Mixture-of-Experts (MoE) architecture is a flexible approach for multi-domain 3D understanding, we observe that conventional feature-only MoE routers may underrepresent local sampling topology under semantic supervision, making expert allocation difficult when semantic consistency coexists with geometric heterogeneity. To overcome this challenge, we propose STAR (Spatial-Topology Aware Routing Framework). Specifically, we introduce a multi-attribute self-supervised pre-training branch, covering topological and textural variations, to anchor cross-domain structural priors. Building upon this, we design a domain-aware expert branch with two mechanisms: Domain-Spatial-Guided Routing (DSR), which captures local topological variations from spatial context, and Entropy-controlled Dynamic Allocation (EDA), which adjusts the number of activated experts according to routing uncertainty. Together, these branches combine stable cross-domain representation learning with adaptive expert allocation. Extensive experiments across various tasks, encompassing both indoor and outdoor scenes, demonstrate the effectiveness of STAR. It achieves 80.1% mIoU on the ScanNet validation set and 77.2% mIoU on S3DIS, consistently improving over strong baselines. Code is available at our project page (https://xmw666.github.io/STAR/).
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。