Geo-EVS: Geometry-Conditioned Extrapolative View Synthesis for Autonomous Driving
Yatong Lan, Rongkui Tang, Lei He · 2026
Read Geo-EVS: Geometry-Conditioned Extrapolative View Synthesis for Autonomous Driving by Yatong Lan, Rongkui Tang, Lei He (2026) 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-2604-07250.txt.
Description
Lan, Tang, and He propose Geo-EVS, a system for synthesizing new camera views outside a car's recorded path. Existing methods degrade off-trajectory because extrapolated poses offer weak geometric support and no dense supervision. Geo-EVS rebuilds colored point clouds and reprojects them to target poses, then trains a diffusion model with artifact masks so it learns to repair missing structure. On Waymo data, it improves sparse-view quality and geometric accuracy most in high-angle, low-coverage settings and helps downstream 3D detection.
Catalog record
| CREATOR | Yatong Lan, Rongkui Tang, Lei He |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 36701 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.07250