Space-Time Forecasting of Dynamic Scenes with Motion-aware Gaussian Grouping
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_ | 1866913084161916928 |
|---|---|
| author | Lee, Junmyeong Choi, Hoseung Cho, Minsu |
| author_facet | Lee, Junmyeong Choi, Hoseung Cho, Minsu |
| contents | Forecasting dynamic scenes remains a fundamental challenge in computer vision, as limited observations make it difficult to capture coherent object-level motion and long-term temporal evolution. We present Motion Group-aware Gaussian Forecasting (MoGaF), a framework for long-term scene extrapolation built upon the 4D Gaussian Splatting representation. MoGaF introduces motion-aware Gaussian grouping and group-wise optimization to enforce physically consistent motion across both rigid and non-rigid regions, yielding spatially coherent dynamic representations. Leveraging this structured space-time representation, a lightweight forecasting module predicts future motion, enabling realistic and temporally stable scene evolution. Experiments on synthetic and real-world datasets demonstrate that MoGaF consistently outperforms existing baselines in rendering quality, motion plausibility, and long-term forecasting stability. Our project page is available at https://slime0519.github.io/mogaf |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_21668 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Space-Time Forecasting of Dynamic Scenes with Motion-aware Gaussian Grouping Lee, Junmyeong Choi, Hoseung Cho, Minsu Computer Vision and Pattern Recognition Graphics Forecasting dynamic scenes remains a fundamental challenge in computer vision, as limited observations make it difficult to capture coherent object-level motion and long-term temporal evolution. We present Motion Group-aware Gaussian Forecasting (MoGaF), a framework for long-term scene extrapolation built upon the 4D Gaussian Splatting representation. MoGaF introduces motion-aware Gaussian grouping and group-wise optimization to enforce physically consistent motion across both rigid and non-rigid regions, yielding spatially coherent dynamic representations. Leveraging this structured space-time representation, a lightweight forecasting module predicts future motion, enabling realistic and temporally stable scene evolution. Experiments on synthetic and real-world datasets demonstrate that MoGaF consistently outperforms existing baselines in rendering quality, motion plausibility, and long-term forecasting stability. Our project page is available at https://slime0519.github.io/mogaf |
| title | Space-Time Forecasting of Dynamic Scenes with Motion-aware Gaussian Grouping |
| topic | Computer Vision and Pattern Recognition Graphics |
| url | https://arxiv.org/abs/2602.21668 |