xxlong0 / Wonder3D

Single Image to 3D using Cross-Domain Diffusion for 3D Generation
https://www.xxlong.site/Wonder3D/
GNU Affero General Public License v3.0
4.83k stars 389 forks source link

Render the SDF from user-defined views #108

Open snowflakewang opened 11 months ago

snowflakewang commented 11 months ago

Hello, your excellent work impressed me a lot. As a novice in 3d vision and graphics, I would like to make some simple exploration based on your project. What I want to do is to render the SDF from user-defined views, i.e. the user gives camera poses and gets multi-view images. If achieving this function, I can make 360-degree demos. But I do not find a renderer implementation in your repo. So could you please give me a brief guide to create such demos as shown in your project page?

Looking forward to your reply. Thank you very very much :)

flamehaze1115 commented 11 months ago

You can check the instant-nsr-pl repo, when you successfully reconstruct the mesh, then you can render arbitrary views. For the diffusion model, it only generate fixed views.