tool · draft
Extend3D
Research model from Seoul National University that generates a town-scale 3D scene from a single image, built on Microsoft's TRELLIS; code released with a CVPR 2026 paper.
What is reusable: open sourceunknown dataAPI unknownunknown participation MITCode on GitHub under MIT; the quick start loads pretrained TRELLIS weights.
The authors are Seungwoo Yoon, Jinmo Kim and Jaesik Park.1 The repository credits Microsoft's TRELLIS, the model its code is built on.1
The quick start takes one image and returns a scene as 3D Gaussians (a scene drawn as many small, soft, coloured blobs) and a mesh.1 The repository shows rendering the Gaussians to a video and exporting the mesh as a GLB file, a common 3D model format.1 A Gradio app gives a browser demo.1
The code is Python, installed locally, under the MIT licence.1,2 Running it takes a Linux x86-64 machine, an NVIDIA GPU with 24 GB of memory, and CUDA 12.4 or later.1 Larger scenes need more memory.1 The quick start loads the pretrained weights named microsoft/TRELLIS-image-large.1
The setup notes say "Coming Soon", and the paper and demo links in the README are not yet live.1,3
- form
- research model
- group
- 3D capture and reconstruction
- homepage
- https://github.com/SNU-VGILab/Extend3D
- licence
- MIT
- cost
- free
Tags: 3d generation, town, single image, gaussian splatting, mesh, computer vision