Published on:
14 March 2024
Primary Category:
Computer Vision and Pattern Recognition
Paper Authors:
Fangfu Liu,
Hanyang Wang,
Weiliang Chen,
Haowen Sun,
Yueqi Duan
Proposes Make-Your-3D for fast subject-specific 3D generation
Uses a co-evolution framework to match desired subject distribution
Employs identity-aware and subject-prior optimization
Works from only a single image in 5 mins, 36x faster than prior work
Produces high-quality and customizable 3D assets
Fast Personalized 3D Model Generation
This paper introduces Make-Your-3D, a novel method to quickly generate high-quality, customizable 3D models of a subject from a single image and text description in just 5 minutes. It works by optimizing and aligning a multi-view diffusion model and 2D generative model to match the distribution of the desired 3D subject.
No comments yet, be the first to start the conversation...
Sign up to comment on this paper