SynCity: Training-Free Generation of 3D Worlds
Paul Engstler *, Aleksandar Shtedritski *, Iro Laina, Christian Rupprecht, Andrea Vedaldi · 2025
Read SynCity: Training-Free Generation of 3D Worlds by Paul Engstler *, Aleksandar Shtedritski *, Iro Laina, Christian Rupprecht, Andrea Vedaldi (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-2503-16420.txt.
Description
Paul Engstler, Aleksandar Shtedritski, Iro Laina, Christian Rupprecht, and Andrea Vedaldi present SynCity in a 2025 paper on building 3D worlds from text. Most 3D generators make single objects, not full places. SynCity makes large scenes with no training and no tuning runs at all.
It blends the shape skill of trained 3D models with the style range of 2D image makers. The world grows tile by tile, each new tile made with its place in the scene in mind. Each tile is then fused into the growing scene before the next one starts. Users get fine control over layout and look as the world expands. The results are rich, varied scenes that feel deep and full.
Catalog record
| CREATOR | Paul Engstler *, Aleksandar Shtedritski *, Iro Laina, Christian Rupprecht, Andrea Vedaldi |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 44 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2503-16420