学习时空一致性用于三维语义场景补全
Learning Spatial-Temporal Consistency for 3D Semantic Scene Completion
把历史帧提升到 3D 场景级占用框架、在时序缓冲中按 2D 可见性与相似性聚合特征，这一思路可直接迁移到任何需要跨帧补全遮挡区域的体素/占用预测任务（如 3D 检测、占据预测）。
原文依据 lifting historical frames into a 3D scene-level occupancy framework, aggregating 2D and 3D historical features from current voxels
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论基于相机的语义场景补全（SSC）能够全面理解整个场景，但由于遮挡和信息不完整，其预测存在歧义。时序 SSC 缓解了这一问题，但现有模型只是简单堆叠多帧时序特征，这会导致几何与语义随时间出现不一致。本文提出 ConSSC，一种学习时空一致性的新型 SSC 方法。其做法是将历史帧提升到三维场景级占用框架中，从当前体素聚合二维与三维历史特征，并在时序缓冲中从二维可见性与相似性线索中学习。具体而言，我们的框架引入两个关键组件：分层体素细化模块，从深度中提取粗略占用并通过体素级表征对其进行细化，从而恢复缺失信息；时序语义聚合模块有效整合来自不同视角与时间点的语义特征，使当前帧中的遮挡区域能够利用历史上下文重建，并将其聚合到相应的体素特征中。无需额外传感器或数据，ConSSC 同时提升了几何与语义一致性。在 SemanticKITTI 与 SSCBench-KITTI-360 数据集上的大量实验表明，ConSSC 在 IoU 与 mIoU 上显著优于最优的基于相机与时序的 SSC 基线。
英文原文摘要 1483 字符
Camera-based Semantic Scene Completion (SSC) is able to comprehensively understand the entire scene, but it suffers from ambiguous predictions due to occlusions and incomplete information. Temporal SSC alleviates this issue, but existing models simply stack multi-frame temporal features, which can lead to inconsistencies between geometry and semantics over time. In this paper, we present ConSSC, a novel SSC method that learns Spatial-Temporal Consistency. It works by lifting historical frames into a 3D scene-level occupancy framework, aggregating 2D and 3D historical features from current voxels, and learning from 2D visibility and similarity cues in a temporal buffer. Specifically, our framework introduces two key components: the Hierarchical Voxel Refinement module, which extracts a coarse occupancy from depth and refines it through voxel-level representations, recovering missing information. The Temporal Semantic Aggregation module effectively integrates semantic features from different viewpoints and time points, enabling the reconstruction of occluded regions in the current frame using historical context, aggregating them into corresponding voxel features. Without additional sensors or data, ConSSC improves both geometric and semantic consistency. Extensive experiments on the SemanticKITTI and SSCBench-KITTI-360 datasets show that ConSSC outperforms state-of-the-art camera-based and temporal SSC baselines by a significant margin in terms of IoU and mIoU.