Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2409.14072 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866915427449307136 |
|---|---|
| author | Zhang, Shuai Wu, Guanjun Xie, Zhoufeng Wang, Xinggang Feng, Bin Liu, Wenyu |
| author_facet | Zhang, Shuai Wu, Guanjun Xie, Zhoufeng Wang, Xinggang Feng, Bin Liu, Wenyu |
| contents | Reconstructing objects and extracting high-quality surfaces play a vital role in the real world. Current 4D representations show the ability to render high-quality novel views for dynamic objects, but cannot reconstruct high-quality meshes due to their implicit or geometrically inaccurate representations. In this paper, we propose a novel representation that can reconstruct accurate meshes from sparse image input, named Dynamic 2D Gaussians (D-2DGS). We adopt 2D Gaussians for basic geometry representation and use sparse-controlled points to capture the 2D Gaussian's deformation. By extracting the object mask from the rendered high-quality image and masking the rendered depth map, we remove floaters that are prone to occur during reconstruction and can extract high-quality dynamic mesh sequences of dynamic objects. Experiments demonstrate that our D-2DGS is outstanding in reconstructing detailed and smooth high-quality meshes from sparse inputs. The code is available at https://github.com/hustvl/Dynamic-2DGS. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2409_14072 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Dynamic 2D Gaussians: Geometrically Accurate Radiance Fields for Dynamic Objects Zhang, Shuai Wu, Guanjun Xie, Zhoufeng Wang, Xinggang Feng, Bin Liu, Wenyu Computer Vision and Pattern Recognition Reconstructing objects and extracting high-quality surfaces play a vital role in the real world. Current 4D representations show the ability to render high-quality novel views for dynamic objects, but cannot reconstruct high-quality meshes due to their implicit or geometrically inaccurate representations. In this paper, we propose a novel representation that can reconstruct accurate meshes from sparse image input, named Dynamic 2D Gaussians (D-2DGS). We adopt 2D Gaussians for basic geometry representation and use sparse-controlled points to capture the 2D Gaussian's deformation. By extracting the object mask from the rendered high-quality image and masking the rendered depth map, we remove floaters that are prone to occur during reconstruction and can extract high-quality dynamic mesh sequences of dynamic objects. Experiments demonstrate that our D-2DGS is outstanding in reconstructing detailed and smooth high-quality meshes from sparse inputs. The code is available at https://github.com/hustvl/Dynamic-2DGS. |
| title | Dynamic 2D Gaussians: Geometrically Accurate Radiance Fields for Dynamic Objects |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2409.14072 |