Advancing high-fidelity 3D and Texture Generation with 2.5D latents

Xin Yang Jiantao Lin Yingjie Xu Haodong Li Ying-Cong Chen · 2025

Read Advancing high-fidelity 3D and Texture Generation with 2.5D latents by Xin Yang Jiantao Lin Yingjie Xu Haodong Li Ying-Cong Chen (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-2505-21050.txt.

Description

Making a 3D object and painting it are usually two separate jobs, and the seams show when shape and surface disagree. Xin Yang and co-authors merge the jobs in a 2025 paper through a halfway representation. Views of an object, its surface angles, and its coordinates are fused into a single latent form that reads as both picture and model.

A big pretrained image model, adapted to this form, dreams up the fused views from text or a photo. A light decoder then lifts the result into a full 3D object. Tests showed cleaner shape-color agreement than staged rivals and strong texture painting on given shapes.

Catalog record

CREATORXin Yang Jiantao Lin Yingjie Xu Haodong Li Ying-Cong Chen
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE38 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2505-21050

Browse the full library · About the Tower of Babel