Synthetic data processing

[Fortunanto]

Hey, as part of a usecase i'm trying out, I'm trying to see how I can train a NeRF from synthetic data, meaning I have complete data of camera location, parameters and everything. i don't want to run colmap or some other sfm tool on those images, because they lack texture, and therefore don't really work in that case. is it possible in any way to support the camera parameters(intrinsics and extrinsics) to nerf studio in the pre-process framework in some way?

Thanks!

[robo-alex]

You might try to organize your data following the conventions.