search

AUTOMATIC1111 / stable-diffusion-webui

Stable Diffusion web UI
GNU Affero General Public License v3.0
141.37k stars 26.72k forks source link

[Feature Request]: Picture to 3D Model #7291

Open pied-stuff opened 1 year ago

pied-stuff commented 1 year ago

What would your feature do ?

Convert a picture to High-Resolution 3D Human Model 3DHuman https://github.com/facebookresearch/pifuhd PIFuHD: Multi-Level Pixel-Aligned Implicit Function for High-Resolution 3D Human Digitization (CVPR 2020)

Proposed workflow

  1. Go to convert to 3D Human Tab
  2. Press convert image after import image
  3. Done

Additional information

No response

ClashSAN commented 1 year ago

Just use this, standalone. https://github.com/openai/point-e turn the notebooks to .py files and run it, also it works on windows with a small change but you can run this on WSL with 4gb vram It needs more vram for the img 2 3d model mode though. But this seems out of scope as built-in

oivio commented 1 year ago

I still would love to see pifuhd as extension. I do believe it would something amazing to have.