lzzcd001/MeshDiffusion
A diffusion model that generates 3D meshes using a direct parameterization of deep marching tetrahedra.
Not currently ranked — collecting fresh signals.
star history
MeshDiffusion implements a score-based generative model for creating 3D meshes. The approach leverages deep marching tetrahedra (DMTet) to directly parameterize and generate 3D shapes. It builds on diffusion model principles but applies them to the 3D geometry domain, enabling high-quality mesh generation across categories like chairs, cars, airplanes, and tables at multiple resolutions.
Frequently asked
- What is lzzcd001/MeshDiffusion?
- A diffusion model that generates 3D meshes using a direct parameterization of deep marching tetrahedra.
- Is MeshDiffusion open source?
- Yes — lzzcd001/MeshDiffusion is open source, released under the MIT license.
- What language is MeshDiffusion written in?
- lzzcd001/MeshDiffusion is primarily written in Python.
- How popular is MeshDiffusion?
- lzzcd001/MeshDiffusion has 834 stars on GitHub.
- Where can I find MeshDiffusion?
- lzzcd001/MeshDiffusion is on GitHub at https://github.com/lzzcd001/MeshDiffusion.