DynSUP: Dynamic Gaussian Splatting from An Unposed Image Pair

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Li, Weihang, Chen, Weirong, Qian, Shenhan, Chen, Jiajie, Cremers, Daniel, Li, Haoang
Formato: Preprint
Publicado: 2024
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866929609670393856
author Li, Weihang
Chen, Weirong
Qian, Shenhan
Chen, Jiajie
Cremers, Daniel
Li, Haoang
author_facet Li, Weihang
Chen, Weirong
Qian, Shenhan
Chen, Jiajie
Cremers, Daniel
Li, Haoang
contents Recent advances in 3D Gaussian Splatting have shown promising results. Existing methods typically assume static scenes and/or multiple images with prior poses. Dynamics, sparse views, and unknown poses significantly increase the problem complexity due to insufficient geometric constraints. To overcome this challenge, we propose a method that can use only two images without prior poses to fit Gaussians in dynamic environments. To achieve this, we introduce two technical contributions. First, we propose an object-level two-view bundle adjustment. This strategy decomposes dynamic scenes into piece-wise rigid components, and jointly estimates the camera pose and motions of dynamic objects. Second, we design an SE(3) field-driven Gaussian training method. It enables fine-grained motion modeling through learnable per-Gaussian transformations. Our method leads to high-fidelity novel view synthesis of dynamic scenes while accurately preserving temporal consistency and object motion. Experiments on both synthetic and real-world datasets demonstrate that our method significantly outperforms state-of-the-art approaches designed for the cases of static environments, multiple images, and/or known poses. Our project page is available at https://colin-de.github.io/DynSUP/.
format Preprint
id arxiv_https___arxiv_org_abs_2412_00851
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle DynSUP: Dynamic Gaussian Splatting from An Unposed Image Pair
Li, Weihang
Chen, Weirong
Qian, Shenhan
Chen, Jiajie
Cremers, Daniel
Li, Haoang
Computer Vision and Pattern Recognition
Recent advances in 3D Gaussian Splatting have shown promising results. Existing methods typically assume static scenes and/or multiple images with prior poses. Dynamics, sparse views, and unknown poses significantly increase the problem complexity due to insufficient geometric constraints. To overcome this challenge, we propose a method that can use only two images without prior poses to fit Gaussians in dynamic environments. To achieve this, we introduce two technical contributions. First, we propose an object-level two-view bundle adjustment. This strategy decomposes dynamic scenes into piece-wise rigid components, and jointly estimates the camera pose and motions of dynamic objects. Second, we design an SE(3) field-driven Gaussian training method. It enables fine-grained motion modeling through learnable per-Gaussian transformations. Our method leads to high-fidelity novel view synthesis of dynamic scenes while accurately preserving temporal consistency and object motion. Experiments on both synthetic and real-world datasets demonstrate that our method significantly outperforms state-of-the-art approaches designed for the cases of static environments, multiple images, and/or known poses. Our project page is available at https://colin-de.github.io/DynSUP/.
title DynSUP: Dynamic Gaussian Splatting from An Unposed Image Pair
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2412.00851