DreamGaussian
Generative 3D Gaussian splatting that makes textured meshes in about 2 minutes.
About
DreamGaussian is an open-source 3D content generation framework by Jiaxiang Tang, Jiawei Ren, Hang Zhou, Ziwei Liu, and Gang Zeng. Instead of slow NeRF-style optimization, it uses a generative 3D Gaussian splatting model with progressive densification, achieving roughly 10x acceleration over prior methods.
It takes a single image or a text prompt and produces a high-quality textured mesh in about 2 minutes, with an efficient algorithm to convert 3D Gaussians into meshes plus UV-space texture refinement. The generated assets can be imported into games and CGI pipelines.
What it can do
- Image-to-3D generation in about 2 minutes
- Text-to-3D generation via score distillation
- Textured mesh extraction with UV texture refinement
- Point cloud-to-3D conversion
Setup
Clone the GitHub repo, install requirements, and run main.py with an image or prompt.
Prerequisites
Python 3, PyTorch, and an NVIDIA GPU with CUDA.
Use cases
Works with
Trust score — 7/10
Scored 7/10 (metrics tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 103; Trending score 72/100 from researcher assessment.
Sources
Every fact above was checked against these sources. Where a claim comes from the provider and couldn't be independently verified, the text says so.