Pose Format and Coordinate Convention in BundleSDF

[engrmusawarali71]

Hello!

Thank you for making this repository public—really appreciate the effort!

I had a quick question regarding the format of the estimated poses provided by BundleSDF. Are the poses represented in the OpenGL or OpenCV coordinate convention?

Specifically, if we plan to train a NeRF model using these estimated poses, what transformations (if any) would you recommend applying to align them correctly with the expected camera convention?

Thanks in advance!