Confusion regarding the colmap example

[AditMeh]

Why does the colmap example invert the rotation matrix here?
https://github.com/nerfstudio-project/viser/blob/567580e0af1ef850b506dc85295e82c53bf55557/examples/04_demos/01_colmap_visualizer.py#L139-L141

The docs state that the rotation needs to be from the parent frame (world) to the local (camera) frame https://viser.studio/main/api/core/scene_api/#viser.SceneApi.add_frame

However, colmap gives you W2C poses in the binary/text file already, so inverting should turn it into C2W poses.
https://colmap.github.io/format.html#images-txt

However, the example works fine and renders properly for my scenes, so there is something I'm missing in my understanding

[brentyi]

Hi!

I think might just be a misreading, from the docs you linked:

rotation to parent frame from local frame (R_pl).

Perhaps the (to, from) wording order is confusing?

This should also be consistent with this doc page: https://viser.studio/main/conventions/#poses