← 研踪 PaperTrailCVF · CVPR 2026:cvf:1d2a7595b585f2b852eb69ec239dbf05 · 记录 #1
Consistent Instance Field for Dynamic Scene Understanding
Wu, JunyiNguyen, Van NguyenPlanche, BenjaminTao, JiachenSun, ChangchangGao, ZhongpaiZhao, ZhenghaoChoudhuri, AnwesaZhang, GengyuZheng, MengWang, FeiranChen, TerrenceYan, YanWu, Ziyan 来源日期（仅到年份） 2026
来源记录日期 2026
雷达观测 2026-11-05 00:55
摘要
We introduce Consistent Instance Field, a continuous and probabilistic spatio-temporal representation for dynamic scene understanding.Unlike prior methods that rely on discrete tracking or view-dependent features, our approach disentangles visibility from persistent object identity by modeling each space-time point with an occupancy probability and a conditional instance distribution. To realize this, we introduce a novel instance-embedded representation based on deformable 3D Gaussians, which jointly encode radiance and semantic information and are learned directly from input RGB images and instance masks through differentiable rasterization.Furthermore, we introduce new mechanisms to calibrate per-Gaussian identities and resample Gaussians toward semantically active regions, ensuring consistent instance representations across space and time. Experiments on HyperNeRF and Neu3D datasets demonstrate that our method significantly outperforms state-of-the-art methods on novel-view panoptic segmentation and open-vocabulary 4D querying tasks.
论文发现出处
这些页面提供论文线索，不代表同行评审或质量认证。
观测到的版本
平台记录号代表本站观测到的元数据变化，不是作者修订号。来源日期仅到月份或年份时，按该月或该年首日存档。
记录 #1 · 2026 · 最新观测版本
Consistent Instance Field for Dynamic Scene Understanding
We introduce Consistent Instance Field, a continuous and probabilistic spatio-temporal representation for dynamic scene understanding.Unlike prior methods that rely on discrete tracking or view-dependent features, our approach disentangles visibility from persistent object identity by modeling each space-time point with an occupancy probability and a conditional instance distribution. To realize this, we introduce a novel instance-embedded representation based on deformable 3D Gaussians, which jointly encode radiance and semantic information and are learned directly from input RGB images and instance masks through differentiable rasterization.Furthermore, we introduce new mechanisms to calibrate per-Gaussian identities and resample Gaussians toward semantically active regions, ensuring consistent instance representations across space and time. Experiments on HyperNeRF and Neu3D datasets demonstrate that our method significantly outperforms state-of-the-art methods on novel-view panoptic segmentation and open-vocabulary 4D querying tasks.
查看此版本解读 →查看此版本原文 ↗