Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Liao, Ting-Hsuan, Liu, Haowen, Xu, Yiran, Ge, Songwei, Yang, Gengshan, Huang, Jia-Bin
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2509.25183
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866912615282769920
author Liao, Ting-Hsuan
Liu, Haowen
Xu, Yiran
Ge, Songwei
Yang, Gengshan
Huang, Jia-Bin
author_facet Liao, Ting-Hsuan
Liu, Haowen
Xu, Yiran
Ge, Songwei
Yang, Gengshan
Huang, Jia-Bin
contents We present PAD3R, a method for reconstructing deformable 3D objects from casually captured, unposed monocular videos. Unlike existing approaches, PAD3R handles long video sequences featuring substantial object deformation, large-scale camera movement, and limited view coverage that typically challenge conventional systems. At its core, our approach trains a personalized, object-centric pose estimator, supervised by a pre-trained image-to-3D model. This guides the optimization of deformable 3D Gaussian representation. The optimization is further regularized by long-term 2D point tracking over the entire input video. By combining generative priors and differentiable rendering, PAD3R reconstructs high-fidelity, articulated 3D representations of objects in a category-agnostic way. Extensive qualitative and quantitative results show that PAD3R is robust and generalizes well across challenging scenarios, highlighting its potential for dynamic scene understanding and 3D content creation.
format Preprint
id arxiv_https___arxiv_org_abs_2509_25183
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle PAD3R: Pose-Aware Dynamic 3D Reconstruction from Casual Videos
Liao, Ting-Hsuan
Liu, Haowen
Xu, Yiran
Ge, Songwei
Yang, Gengshan
Huang, Jia-Bin
Computer Vision and Pattern Recognition
We present PAD3R, a method for reconstructing deformable 3D objects from casually captured, unposed monocular videos. Unlike existing approaches, PAD3R handles long video sequences featuring substantial object deformation, large-scale camera movement, and limited view coverage that typically challenge conventional systems. At its core, our approach trains a personalized, object-centric pose estimator, supervised by a pre-trained image-to-3D model. This guides the optimization of deformable 3D Gaussian representation. The optimization is further regularized by long-term 2D point tracking over the entire input video. By combining generative priors and differentiable rendering, PAD3R reconstructs high-fidelity, articulated 3D representations of objects in a category-agnostic way. Extensive qualitative and quantitative results show that PAD3R is robust and generalizes well across challenging scenarios, highlighting its potential for dynamic scene understanding and 3D content creation.
title PAD3R: Pose-Aware Dynamic 3D Reconstruction from Casual Videos
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2509.25183