GS-Surrogate: Deformable Gaussian Splatting for Parameter Space Exploration of Ensemble Simulations
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2026
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866908945551982592 |
|---|---|
| author | Li, Ziwei Perera, Rumali Forbes, Angus Moreland, Ken Pugmire, Dave Klasky, Scott Chao, Wei-Lun Shen, Han-Wei |
| author_facet | Li, Ziwei Perera, Rumali Forbes, Angus Moreland, Ken Pugmire, Dave Klasky, Scott Chao, Wei-Lun Shen, Han-Wei |
| contents | Exploring ensemble simulations is increasingly important across many scientific domains. However, supporting flexible post-hoc exploration remains challenging due to the trade-off between storing the expensive raw data and flexibly adjusting visualization settings. Existing visualization surrogate models have improved this workflow, but they either operate in image space without an explicit 3D representation or rely on neural radiance fields that are computationally expensive for interactive exploration and encode all parameter-driven variations within a single implicit field. In this work, we introduce GS-Surrogate, a deformable Gaussian Splatting-based visualization surrogate for parameter-space exploration. Our method first constructs a canonical Gaussian field as a base 3D representation and adapts it through sequential parameter-conditioned deformations. By separating simulation-related variations from visualization-specific changes, this explicit formulation enables efficient and controllable adaptation to different visualization tasks, such as isosurface extraction and transfer function editing. We evaluate our framework on a range of simulation datasets, demonstrating that GS-Surrogate enables real-time and flexible exploration across both simulation and visualization parameter spaces. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_06358 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | GS-Surrogate: Deformable Gaussian Splatting for Parameter Space Exploration of Ensemble Simulations Li, Ziwei Perera, Rumali Forbes, Angus Moreland, Ken Pugmire, Dave Klasky, Scott Chao, Wei-Lun Shen, Han-Wei Graphics Artificial Intelligence Exploring ensemble simulations is increasingly important across many scientific domains. However, supporting flexible post-hoc exploration remains challenging due to the trade-off between storing the expensive raw data and flexibly adjusting visualization settings. Existing visualization surrogate models have improved this workflow, but they either operate in image space without an explicit 3D representation or rely on neural radiance fields that are computationally expensive for interactive exploration and encode all parameter-driven variations within a single implicit field. In this work, we introduce GS-Surrogate, a deformable Gaussian Splatting-based visualization surrogate for parameter-space exploration. Our method first constructs a canonical Gaussian field as a base 3D representation and adapts it through sequential parameter-conditioned deformations. By separating simulation-related variations from visualization-specific changes, this explicit formulation enables efficient and controllable adaptation to different visualization tasks, such as isosurface extraction and transfer function editing. We evaluate our framework on a range of simulation datasets, demonstrating that GS-Surrogate enables real-time and flexible exploration across both simulation and visualization parameter spaces. |
| title | GS-Surrogate: Deformable Gaussian Splatting for Parameter Space Exploration of Ensemble Simulations |
| topic | Graphics Artificial Intelligence |
| url | https://arxiv.org/abs/2604.06358 |