TGSFormer：面向具身语义场景补全的可扩展时序高斯泼溅
TGSFormer: Scalable Temporal Gaussian Splatting for Embodied Semantic Scene Completion
用持久化的高斯记忆加置信度感知的跨注意力做时序融合，替代依赖帧缓存与图像一致性的做法，可复用到任何需要在线增量重建、且场景尺度无界的具身感知或流式建图任务中。
原文依据 a Dual Temporal Encoder jointly processes current and historical Gaussian features through confidence-aware cross-attention
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论具身三维语义场景补全（SSC）从连续的第一视角观测中推断稠密几何与语义。现有大多数基于高斯的方法依赖在预定义空间范围内随机初始化大量图元，导致冗余，并且对无界场景的可扩展性很差。近期的深度引导方法缓解了该问题，但仍局限于局部，随着规模增大而出现延迟与内存开销。为克服这些挑战，我们提出 TGSFormer，一个面向具身 SSC 的可扩展时序高斯泼溅框架。它维护一个持久的高斯记忆用于时序预测，不依赖图像一致性或帧缓存。在时序融合方面，双时序编码器通过置信度感知的交叉注意力联合处理当前与历史高斯特征。随后，置信度感知体素融合模块将重叠图元合并为体素对齐的表示，从而调控密度并保持紧凑性。大量实验表明，TGSFormer 在局部与具身 SSC 基准上均取得最优结果，在显著更少图元的情况下提供更优的精度与可扩展性，同时保持长期场景完整性的一致性。代码将在论文被接收后发布。
英文原文摘要 1294 字符
Embodied 3D Semantic Scene Completion (SSC) infers dense geometry and semantics from continuous egocentric observations. Most existing Gaussian-based methods rely on random initialization of many primitives within predefined spatial bounds, resulting in redundancy and poor scalability to unbounded scenes. Recent depth-guided approach alleviates this issue but remains local, suffering from latency and memory overhead as scale increases. To overcome these challenges, we propose TGSFormer, a scalable Temporal Gaussian Splatting framework for embodied SSC. It maintains a persistent Gaussian memory for temporal prediction, without relying on image coherence or frame caches. For temporal fusion, a Dual Temporal Encoder jointly processes current and historical Gaussian features through confidence-aware cross-attention. Subsequently, a Confidence-aware Voxel Fusion module merges overlapping primitives into voxel-aligned representations, regulating density and maintaining compactness. Extensive experiments demonstrate that TGSFormer achieves state-of-the-art results on both local and embodied SSC benchmarks, offering superior accuracy and scalability with significantly fewer primitives while maintaining consistent long-term scene integrity. The code will be released upon acceptance.