RiGS: Rigid-aware 4D Gaussian Splatting from a Single Monocular Video

Chenyu Wu, Wanhua Li, Zhu-Tian Chen, Hanspeter Pfister · 2026

Read RiGS: Rigid-aware 4D Gaussian Splatting from a Single Monocular Video by Chenyu Wu, Wanhua Li, Zhu-Tian Chen, Hanspeter Pfister (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-23672.txt.

Description

Wu and colleagues present RiGS, rigid-aware 4D Gaussian splatting from one video. Existing methods lose temporal consistency or miss high-frequency dynamics. RiGS uses three Gaussian types: static for backgrounds, rigid for long-term motion, and transient for short-term dynamics. An object-wise dynamic mask aggregates long-range motion and guides the static-dynamic split. Rigid Gaussians can become transient ones by duration, and both train under scene flow guidance for dense 3D motion supervision. RiGS reaches top performance on novel view synthesis benchmarks.

Catalog record

CREATORChenyu Wu, Wanhua Li, Zhu-Tian Chen, Hanspeter Pfister
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE50296
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.23672

Browse the full library · About the Tower of Babel