Tactile DreamFusion: Exploiting Tactile Sensing for 3D Generation

Ruihan Gao, Kangle Deng, Gengshan Yang, Wenzhen Yuan, Jun-Yan Zhu · 2024

Read Tactile DreamFusion: Exploiting Tactile Sensing for 3D Generation by Ruihan Gao, Kangle Deng, Gengshan Yang, Wenzhen Yuan, Jun-Yan Zhu (2024) 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-2412-06785.txt.

Description

Ruihan Gao and co-authors add the sense of touch to 3D content generation in this 2024 study. Image-driven 3D tools make pretty objects, but their surfaces often come out too smooth, or fine detail gets painted on as color instead of real shape. The team argues that touch data, which records true surface geometry, can fix that gap between looks and feel.

Their method builds a light 3D texture field that produces matching visual and tactile textures, guided by image models trained on both kinds of data. Fine touch readings steer the visual detail, and a custom-tuned model refines touch patches. A multi-part pipeline lets each region of an object carry its own texture. Tests from both text and image prompts showed sharper, more faithful fine geometry.

Catalog record

CREATORRuihan Gao, Kangle Deng, Gengshan Yang, Wenzhen Yuan, Jun-Yan Zhu
YEAR2024
COLLECTIONPAPERS
FORMATTXT
FILE SIZE45 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2412-06785

Browse the full library · About the Tower of Babel