[ Back ]

Strips as Tokens: Artist Mesh Generation with Native UV Segmentation

Strips as Tokens: Artist Mesh Generation with Native UV Segmentation

Rui Xu, Dafei Qin, Kaichun Qiao, Qiujie Dong, Huaijin Pi, Qixuan Zhang, Longwen Zhang, Lan Xu, Jingyi Yu, Wenping Wang, Taku Komura

SIGGRAPH 2026 Journal Track (Conditionally Accepted) ยท 2026

Abstract

Strips as Tokens (SATO) is an autoregressive mesh generation framework that uses a token ordering inspired by triangle strips. By explicitly encoding UV boundaries in a connected face sequence, it preserves the edge flow and semantic structure expected in artist-created meshes, while supporting unified decoding to both triangle and quadrilateral meshes.

Imported from Rui Xu’s publication page.