Disentangled 4D Gaussian Splatting: Rendering High-Resolution Dynamic World at 343 FPS
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866908619407097856 |
|---|---|
| author | Feng, Hao Sun, Hao Xie, Wei Zuo, Zhi Liu, Zhengzhe |
| author_facet | Feng, Hao Sun, Hao Xie, Wei Zuo, Zhi Liu, Zhengzhe |
| contents | While dynamic novel view synthesis from 2D videos has seen progress, achieving efficient reconstruction and rendering of dynamic scenes remains a challenging task. In this paper, we introduce Disentangled 4D Gaussian Splatting (Disentangled4DGS), a novel representation and rendering pipeline that achieves real-time performance without compromising visual fidelity. Disentangled4DGS decouples the temporal and spatial components of 4D Gaussians, avoiding the need for slicing first and four-dimensional matrix calculations in prior methods. By projecting temporal and spatial deformations into dynamic 2D Gaussians and deferring temporal processing, we minimize redundant computations of 4DGS. Our approach also features a gradient-guided flow loss and temporal splitting strategy to reduce artifacts. Experiments demonstrate a significant improvement in rendering speed and quality, achieving 343 FPS when render 1352*1014 resolution images on a single RTX3090 while reducing storage requirements by at least 4.5%. Our approach sets a new benchmark for dynamic novel view synthesis, outperforming existing methods on both multi-view and monocular dynamic scene datasets. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2503_22159 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Disentangled 4D Gaussian Splatting: Rendering High-Resolution Dynamic World at 343 FPS Feng, Hao Sun, Hao Xie, Wei Zuo, Zhi Liu, Zhengzhe Graphics Computer Vision and Pattern Recognition While dynamic novel view synthesis from 2D videos has seen progress, achieving efficient reconstruction and rendering of dynamic scenes remains a challenging task. In this paper, we introduce Disentangled 4D Gaussian Splatting (Disentangled4DGS), a novel representation and rendering pipeline that achieves real-time performance without compromising visual fidelity. Disentangled4DGS decouples the temporal and spatial components of 4D Gaussians, avoiding the need for slicing first and four-dimensional matrix calculations in prior methods. By projecting temporal and spatial deformations into dynamic 2D Gaussians and deferring temporal processing, we minimize redundant computations of 4DGS. Our approach also features a gradient-guided flow loss and temporal splitting strategy to reduce artifacts. Experiments demonstrate a significant improvement in rendering speed and quality, achieving 343 FPS when render 1352*1014 resolution images on a single RTX3090 while reducing storage requirements by at least 4.5%. Our approach sets a new benchmark for dynamic novel view synthesis, outperforming existing methods on both multi-view and monocular dynamic scene datasets. |
| title | Disentangled 4D Gaussian Splatting: Rendering High-Resolution Dynamic World at 343 FPS |
| topic | Graphics Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2503.22159 |