Closed mmahdavian closed 4 weeks ago
This is because we follow MapTR's choice of coordinate frame so that the comparison can be fair (under the same coordinate system). This was explained in Sec4.1 of the paper.
MapTR and many other previous works use the lidar coordinate system because they have models fusing camera inputs and lidar inputs.
Hi. I was wondering what was the reason that you converted ego coordinate frame to lidar? Why didn't you use ego coordinate frame for e2g_translation and e2g_rotation?
Thank You