Synthesizing Consistent Novel Views via 3D Epipolar Attention without Re-Training
Botao Ye, Sifei Liu, Xueting Li, Marc Pollefeys, Ming-Hsuan Yang · 2025
Read Synthesizing Consistent Novel Views via 3D Epipolar Attention without Re-Training by Botao Ye, Sifei Liu, Xueting Li, Marc Pollefeys, Ming-Hsuan Yang (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2502-18219.txt.
Description
Botao Ye, Sifei Liu, Xueting Li, Marc Pollefeys, and Ming-Hsuan Yang present a method for novel view synthesis in a 2025 arXiv paper. Large diffusion models can create new views from a single image, but the new views often clash with the input view. The clash comes from poor use of context in the reference image.
Their fix uses epipolar geometry to find and fetch the overlapping parts of the input view. Those parts guide the creation of each target view, with no training or learnable parts. The method extends to many views at once. Tests show more consistent views and better 3D reconstruction downstream.
Catalog record
| CREATOR | Botao Ye, Sifei Liu, Xueting Li, Marc Pollefeys, Ming-Hsuan Yang |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 53 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2502-18219