search

ashawkey / stable-dreamfusion

Text-to-3D & Image-to-3D & Mesh Exportation with NeRF + Diffusion.
Apache License 2.0
8.28k stars 730 forks source link

The texture when I use fine-tuning with DMTet (mesh representation) #216

Closed ug-kim closed 1 year ago

ug-kim commented 1 year ago

I have some questions about DMTET fine-tuning.

DMTet makes mesh as final representations; I'm curious how we learn the texture of mesh at DMTet fine-tuning stages.

And how do you render mesh rendering with geometry and texture?

It would be appreciated to tell me about the method and code part.

Thanks for sharing your awesome work.

ashawkey commented 1 year ago

@ug-kim Hi, the code of DMTet is majorly adapted from nvdiffrec. You may check their paper for a detailed explanation. In summary, we still use the 3D hashgrid from NeRF stage to encode the texture during training, and only bake it into 2D texture images at exporting.

ug-kim commented 1 year ago

@ashawkey Thank you for your fast and kind response! I will check the paper you mentioned :)