It’s quite impressive that they can incorporate that final camera that is presumably user-mounted. Things like this can be “easy” if you know the angles and positions of all of the cameras very precisely, but I think that having that final camera in an “unknown” position really complicates things. I wonder if it can…