Guardado en:
Detalles Bibliográficos
Autores principales: Yuan, Yu, Wickremasinghe, Tharindu, Nadir, Zeeshan, Wang, Xijun, Chi, Yiheng, Chan, Stanley H.
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:https://arxiv.org/abs/2512.03350
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866917365319467008
author Yuan, Yu
Wickremasinghe, Tharindu
Nadir, Zeeshan
Wang, Xijun
Chi, Yiheng
Chan, Stanley H.
author_facet Yuan, Yu
Wickremasinghe, Tharindu
Nadir, Zeeshan
Wang, Xijun
Chi, Yiheng
Chan, Stanley H.
contents Images and videos are discrete 2D projections of the 4D world (3D space + time). Most visual understanding, prediction, and generation operate directly on 2D observations, leading to suboptimal performance. We propose SeeU, a novel approach that learns the continuous 4D dynamics and generate the unseen visual contents. The principle behind SeeU is a new 2D$\to$4D$\to$2D learning framework. SeeU first reconstructs the 4D world from sparse and monocular 2D frames (2D$\to$4D). It then learns the continuous 4D dynamics on a low-rank representation and physical constraints (discrete 4D$\to$continuous 4D). Finally, SeeU rolls the world forward in time, re-projects it back to 2D at sampled times and viewpoints, and generates unseen regions based on spatial-temporal context awareness (4D$\to$2D). By modeling dynamics in 4D, SeeU achieves continuous and physically-consistent novel visual generation, demonstrating strong potentials in multiple tasks including unseen temporal generation, unseen spatial generation, and video editing. All data and code will be public at https://yuyuanspace.com/SeeU/
format Preprint
id arxiv_https___arxiv_org_abs_2512_03350
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle SeeU: Seeing the Unseen World via 4D Dynamics-aware Generation
Yuan, Yu
Wickremasinghe, Tharindu
Nadir, Zeeshan
Wang, Xijun
Chi, Yiheng
Chan, Stanley H.
Computer Vision and Pattern Recognition
Images and videos are discrete 2D projections of the 4D world (3D space + time). Most visual understanding, prediction, and generation operate directly on 2D observations, leading to suboptimal performance. We propose SeeU, a novel approach that learns the continuous 4D dynamics and generate the unseen visual contents. The principle behind SeeU is a new 2D$\to$4D$\to$2D learning framework. SeeU first reconstructs the 4D world from sparse and monocular 2D frames (2D$\to$4D). It then learns the continuous 4D dynamics on a low-rank representation and physical constraints (discrete 4D$\to$continuous 4D). Finally, SeeU rolls the world forward in time, re-projects it back to 2D at sampled times and viewpoints, and generates unseen regions based on spatial-temporal context awareness (4D$\to$2D). By modeling dynamics in 4D, SeeU achieves continuous and physically-consistent novel visual generation, demonstrating strong potentials in multiple tasks including unseen temporal generation, unseen spatial generation, and video editing. All data and code will be public at https://yuyuanspace.com/SeeU/
title SeeU: Seeing the Unseen World via 4D Dynamics-aware Generation
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2512.03350