VOXL2 object detection and tracking

[lincior]

Hi, I'm planning to buy one of the ModalAI development drones for the purpose of following a target object that is detected via neural networks. I need to understand if this is achievable out of the box with VOXL2.
I understand that voxl-tflite-server can be used for inference. What I don't understand is whether I can to pass the detection output to some other module of the MPA, so that (1) detection -> tracking and (2) tracking -> control.
In other words, pretend what I'm trying to achieve is to have the drone following me autonomously while I walk.

Regarding transforming detections into traces for tracking, I saw this link has a reference to something that could look like object tracking, which I expect might be used also for tracking april tags, but I can't figure out the docs which seem incomplete. Can anyone point me to useful resources? Should I just use opencv instead?

Regarding passing tracking information to the system so that it follows it (control modules) I am nowhere: can someone help me to get started here?