FreeGaussian: Annotation-free Control of Articulated Objects via 3D Gaussian Splats with Flow Derivatives

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Chen, Qizhi, Qu, Delin, Liu, Junli, Tang, Yiwen, Song, Haoming, Wang, Dong, Yuan, Yuan, Zhao, Bin
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910054906593280
author Chen, Qizhi
Qu, Delin
Liu, Junli
Tang, Yiwen
Song, Haoming
Wang, Dong
Yuan, Yuan
Zhao, Bin
author_facet Chen, Qizhi
Qu, Delin
Liu, Junli
Tang, Yiwen
Song, Haoming
Wang, Dong
Yuan, Yuan
Zhao, Bin
contents Reconstructing controllable Gaussian splats for articulated objects from monocular video is especially challenging due to its inherently insufficient constraints. Existing methods address this by relying on dense masks and manually defined control signals, limiting their real-world applications. In this paper, we propose an annotation-free method, FreeGaussian, which mathematically disentangles camera egomotion and articulated movements via flow derivatives. By establishing a connection between 2D flows and 3D Gaussian dynamic flow, our method enables optimization and continuity of dynamic Gaussian motions from flow priors without any control signals. Furthermore, we introduce a 3D spherical vector controlling scheme, which represents the state as a 3D Gaussian trajectory, thereby eliminating the need for complex 1D control signal calculations and simplifying controllable Gaussian modeling. Extensive experiments on articulated objects demonstrate the state-of-the-art visual performance and precise, part-aware controllability of our method. Code is available at: https://github.com/Tavish9/freegaussian.
format Preprint
id arxiv_https___arxiv_org_abs_2410_22070
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle FreeGaussian: Annotation-free Control of Articulated Objects via 3D Gaussian Splats with Flow Derivatives
Chen, Qizhi
Qu, Delin
Liu, Junli
Tang, Yiwen
Song, Haoming
Wang, Dong
Yuan, Yuan
Zhao, Bin
Computer Vision and Pattern Recognition
Machine Learning
Reconstructing controllable Gaussian splats for articulated objects from monocular video is especially challenging due to its inherently insufficient constraints. Existing methods address this by relying on dense masks and manually defined control signals, limiting their real-world applications. In this paper, we propose an annotation-free method, FreeGaussian, which mathematically disentangles camera egomotion and articulated movements via flow derivatives. By establishing a connection between 2D flows and 3D Gaussian dynamic flow, our method enables optimization and continuity of dynamic Gaussian motions from flow priors without any control signals. Furthermore, we introduce a 3D spherical vector controlling scheme, which represents the state as a 3D Gaussian trajectory, thereby eliminating the need for complex 1D control signal calculations and simplifying controllable Gaussian modeling. Extensive experiments on articulated objects demonstrate the state-of-the-art visual performance and precise, part-aware controllability of our method. Code is available at: https://github.com/Tavish9/freegaussian.
title FreeGaussian: Annotation-free Control of Articulated Objects via 3D Gaussian Splats with Flow Derivatives
topic Computer Vision and Pattern Recognition
Machine Learning
url https://arxiv.org/abs/2410.22070