how can i get 3D bounding boxes from the GT camera?
I assume you’re referring to the AURELION ground truth camera? That outputs a 2D bounding box, not 3D. If you need 3D bounding box ground truth data, I would recommend using the OSI GT data for that. Are you using the VESI 1.0 or 2.0? The demo package for VESI 1.0 includes a deserializeOSIGT block which will output the GT data for moving and stationary objects.