Paper Title:
Make-it-Real: Unleashing Large Multimodal Model's Ability for Painting 3D Objects with Realistic Materials
Published on:
25 April 2024
Primary Category:
Computer Vision and Pattern Recognition
Paper Authors:
Ye Fang,
Zeyi Sun,
Tong Wu,
Jiaqi Wang,
Ziwei Liu,
Gordon Wetzstein,
Dahua Lin
Make-it-Real pipeline assigns materials to 3D objects using GPT-4V's recognition capabilities
A detailed material library with thousands of entries and descriptions is constructed
Visual and hierarchical text prompts guide precise material matching for object components
Matched materials become references to generate realistic SVBRDF maps
The approach significantly enhances visual realism and integrates into 3D workflows
Realistic Material Assignment for 3D Objects
This paper presents Make-it-Real, a novel approach that leverages large language models to identify materials from images and assign them to 3D objects. This allows creating realistic material properties for existing 3D assets and generated models.
No comments yet, be the first to start the conversation...
Sign up to comment on this paper