EX-4D: EXtreme Viewpoint 4D Video Synthesis via Depth Watertight Mesh
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866910991568076800 |
|---|---|
| author | Hu, Tao Peng, Haoyang Liu, Xiao Ma, Yuewen |
| author_facet | Hu, Tao Peng, Haoyang Liu, Xiao Ma, Yuewen |
| contents | Generating high-quality camera-controllable videos from monocular input is a challenging task, particularly under extreme viewpoint. Existing methods often struggle with geometric inconsistencies and occlusion artifacts in boundaries, leading to degraded visual quality. In this paper, we introduce EX-4D, a novel framework that addresses these challenges through a Depth Watertight Mesh representation. The representation serves as a robust geometric prior by explicitly modeling both visible and occluded regions, ensuring geometric consistency in extreme camera pose. To overcome the lack of paired multi-view datasets, we propose a simulated masking strategy that generates effective training data only from monocular videos. Additionally, a lightweight LoRA-based video diffusion adapter is employed to synthesize high-quality, physically consistent, and temporally coherent videos. Extensive experiments demonstrate that EX-4D outperforms state-of-the-art methods in terms of physical consistency and extreme-view quality, enabling practical 4D video generation. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2506_05554 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | EX-4D: EXtreme Viewpoint 4D Video Synthesis via Depth Watertight Mesh Hu, Tao Peng, Haoyang Liu, Xiao Ma, Yuewen Computer Vision and Pattern Recognition Generating high-quality camera-controllable videos from monocular input is a challenging task, particularly under extreme viewpoint. Existing methods often struggle with geometric inconsistencies and occlusion artifacts in boundaries, leading to degraded visual quality. In this paper, we introduce EX-4D, a novel framework that addresses these challenges through a Depth Watertight Mesh representation. The representation serves as a robust geometric prior by explicitly modeling both visible and occluded regions, ensuring geometric consistency in extreme camera pose. To overcome the lack of paired multi-view datasets, we propose a simulated masking strategy that generates effective training data only from monocular videos. Additionally, a lightweight LoRA-based video diffusion adapter is employed to synthesize high-quality, physically consistent, and temporally coherent videos. Extensive experiments demonstrate that EX-4D outperforms state-of-the-art methods in terms of physical consistency and extreme-view quality, enabling practical 4D video generation. |
| title | EX-4D: EXtreme Viewpoint 4D Video Synthesis via Depth Watertight Mesh |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2506.05554 |