핵심 개념
Automated texturing method using Stable Diffusion depth enhances quality and speed of 3D models.
초록
The paper introduces EucliDreamer, a novel method for generating textures for 3D models using text prompts and 3D meshes. By incorporating Stable Diffusion depth within the Score Distillation Sampling (SDS) process, the model produces high-quality textures in various art styles faster than existing methods. The research addresses common issues in 3D texturing like wrong semantics, view inconsistency, excessive shadows, and bad color tones. Through experiments and user studies, EucliDreamer demonstrates significant improvements in texture generation quality, diversity, and consistency across different viewpoints.
통계
Global market size of 3D modeling and animation: $6 billion [12]
Over 800K textured models in Objaverse dataset [7]
EucliDreamer converges in around 4300 steps [5]
인용구
"Our model can generate more satisfactory results and produce various art styles for the same object." - EucliDreamer Paper
"EucliDreamer achieves faster time when generating textures of comparable quality." - EucliDreamer Paper