Questions about the precompute_pose_gt.py about md4all HOT 2 CLOSED

Hi @Amor-ZYF,

Thank you for pointing that out! You are right, that was indeed wrong. However, we double-checked and want to clarify that this does not affect our project: it does not affect any of the results we reported in the paper.
In fact, we do not use the 'next' timestamp and the 'prev' timestamp from the 'poses_synchronized.json' file for the RobotCar related code within our project. We initially included those attributes, but we eventually did not make use of them, so the index swapping causes no harm.

As our code does not depend on such information, we updated the precompute_pose_gt.py script and uploaded the updated 'poses_synchronized.json' to the Google drive for both scenes such that the 'prev' and 'next' timestamps are excluded (as they are anyways not used). If your own code modifications depend on those attributes, then you can fix the line of code, as you suggested, by using timestamp[1] for the key 'next'.

Thanks again, and have a nice week!

from md4all.

Hi! Closing this for inactivity. Feel free to reopen it if you need to.

from md4all.