Open bxcharlie opened 1 year ago
Hello,
I'm not sure if I understand the question. The camera matrix is computed here https://github.com/NVlabs/nvdiffrec/blob/main/dataset/dataset_llff.py#L85
from data read from the file poses_bounds.npy
, in case of an LLFF dataset.
This file https://github.com/NVlabs/nvdiffrec/blob/main/dataset/dataset_mesh.py shows a case where we generate (random) poses on the fly.
When I use the dataset_llff, how to change the camera pose to adapt other render method that changed the camera pose.