REArtGS++: Generalizable Articulation Reconstruction with Temporal Geometry Constraint via Planar Gaussian Splatting
Di Wu, Liu Liu, Anran Huang, Yuyan Liu, Qiaoyu Jun, Shaofan Liu, Liangtu Song, Cewu Lu
Embedded Files
Abstract
Abstract
Articulated objects are pervasive in daily environments, such as drawers and refrigerators. Towards their part-level surface reconstruction and joint parameter estimation, REArtGS introduces a category-agnostic approach using multi-view RGB images at two different states. However, we observe that REArtGS still struggles with screw-joint or multi-part objects and lacks geometric constraints for unseen states. In this paper, we propose REArtGS++, a novel method towards generalizable articulated object reconstruction with temporal geometry constraint and planar Gaussian splatting. We first model a decoupled screw motion for each joint without type prior, and jointly optimize part-aware Gaussians with joint parameters through part motion blending. To introduce time-continuous geometric constraint for articulated modeling, we encourage Gaussians to be planar and propose a temporally consistent regularization between planar normal and depth through Taylor first-order expansion. Extensive experiments on both synthetic and real-world articulated objects demonstrate our superiority in generalizable part-level surface reconstruction and joint parameter estimation, compared to existing approaches. Codes will be made publicly available.
Video
Framework
Framework
Qualitative Results of Two-Part Articulated Objects on Paris Dataset
Qualitative Results of Multi-Part Articulated Objects on ArtGS-Multi Dataset
Qualitative Results of Screw-Joint Articulated Objects on PartNet-Mobility Dataset
Qualitative Results of Real-World Articulated ObjectsÂ
Rendering Results Using Multi-View RGB Images at Two States
Two-Part Articulated Objects
Screw-Joint Articulated Objects
Multi-Part Articulated Objects
Real-World Articulated Objects
Page updated
Google Sites
Report abuse