Sketch with 3D with ComyUI
The generation of 3D models with AI holds great potential for form finding and designing 3D assets. Particularly the ability being able to quickly sketch an idea and get a 3D representation within an environment can be a very useful workflow especially as this technology becomes more accurate.
For this tutorial I will be using TripoSR by stability AI for Fast 3D Object Reconstruction from a Single Image . Only three nodes are actually needed for this workflow and has similar results to much more complicated workflows or even paid services.
00:00 Intro
00:55 TripoSR install
01:48 Sketch to Image workflow
02:14 TripoSR nodes
02:52 Background remover
04:30 3D generation in comfyui
05:10 Changing input
06:34 Google 3D Tiles
07:13 3Ds Max
07:35 CRM 2d to 3d model
07:42 Outro
Install: • One-click installs for Open Source AI tool...
Part I: • ComfyUI: Getting started for Stable Diffus...
Part II: • ComfyUI: Intro to Masking and In-painting ...
Part III: • ComfyUI: Intro to Control Net- Sketch to R...
Part IV: • Best Photorealistic Models with ComfyUI
TripoSR Models, github and academic papers:
https://huggingface.co/stabilityai/Tr...
https://github.com/flowtyone/ComfyUI-...
https://arxiv.org/abs/2403.02151
https://github.com/VAST-AI-Research/T...
CRM Models, github and academic papers:
https://ml.cs.tsinghua.edu.cn/~zhengy...
https://huggingface.co/Zhengyi/CRM
Other 2D to 3D workflows:
Shape-E: GitHub - openai/shap-e: Generate 3D objects conditioned on text or images
Cat3D: CAT3D: Create Anything in 3D with Multi-View Diffusion Models
Instagram account: / urbandecoders
#ai #architecture #generativedesign #parametric #stablediffusion #upscaling #animation #comfui #image #design #generativedesign #painting #sketchtorender #styletransfer #sketchtorender