By my understanding that isn't usually the job of the camera itself, they don't have the processing power to do it at a reasonable speed, so I wouldn't expect this firmware to offer it.
That sort of task is usually done by the software that reads from the cameras, be that something local like https://frigate.video/ or the cloud-based ML of the manufacturer of your camera.
That sort of task is usually done by the software that reads from the cameras, be that something local like https://frigate.video/ or the cloud-based ML of the manufacturer of your camera.