ReScene4D：演化室内三维场景的时序一致语义实例分割
ReScene4D: Temporally Consistent Semantic Instance Segmentation of Evolving Indoor 3D Scenes
针对间歇性采集的室内 3D 扫描，提出时序稀疏 4D 语义实例分割任务，通过时空对比损失和序列化实现跨观测的实例身份关联，可直接用于动态室内场景的长期建图与变化检测。
原文依据 We introduce and formalize the task of temporally sparse 4D indoor semantic instance segmentation (SIS)
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论室内环境会随着物体的移动、出现或离开而演变。要捕捉这些动态，需要在间歇性采集的三维扫描之间保持时序一致的实例身份，即使变化未被观测到。我们引入并形式化了时序稀疏的四维室内语义实例分割（SIS）任务，该任务联合分割、识别并时序关联物体实例。这一设定对现有三维语义实例分割方法构成挑战，因为它们缺乏时序推理，需要离散的匹配步骤；同时也对四维激光雷达方法构成挑战，因为这些方法依赖高频时序测量，而这在室内环境的长期演变中并不常见，导致性能不佳。我们提出 ReScene4D，一种新颖的方法，能够将三维语义实例分割架构适配到四维语义实例分割，而无需密集观测。我们的方法通过时空对比损失、掩码和序列化实现时序信息共享，以自适应地利用跨观测的几何和语义先验。这种共享上下文实现了一致的实例跟踪，并提升了标准三维语义实例分割的性能。为了评估该任务，我们定义了一个新指标 t-mAP，它扩展了平均精度均值以奖励时序身份一致性。ReScene4D 在 3RScan 数据集上取得了最优性能，为理解演化中的室内场景建立了新的基准。
英文原文摘要 1401 字符
Indoor environments evolve as objects move, appear, or leave the scene. Capturing these dynamics requires maintaining temporally consistent instance identities across intermittently captured 3D scans, even when changes are unobserved. We introduce and formalize the task of temporally sparse 4D indoor semantic instance segmentation (SIS), which jointly segments, identifies, and temporally associates object instances. This setting poses a challenge for existing 3DSIS methods, which require a discrete matching step due to their lack of temporal reasoning, and for 4D LiDAR approaches, which perform poorly due to their reliance on high-frequency temporal measurements that are uncommon in the longer-horizon evolution of indoor environments. We propose ReScene4D, a novel method that adapts 3DSIS architectures for 4DSIS without needing dense observations. Our method enables temporal information sharing--using spatiotemporal contrastive loss, masking, and serialization--to adaptively leverage geometric and semantic priors across observations. This shared context enables consistent instance tracking and improves standard 3DSIS performance. To evaluate this task, we define a new metric, t-mAP, that extends mAP to reward temporal identity consistency. ReScene4D achieves state-of-the-art performance on the 3RScan dataset, establishing a new benchmark for understanding evolving indoor scenes.