Identify persons from a distance of two meters

I want to identify people who come close to my device then the device perform a specific operation based on each person
Available ways
1- image recognition (heavy for openmv)
2-qr code detection (low range)
3- apriltag (it is not possible to write data on it)

Is there another solution to this?

Hi, we’re going to release bounding box CNN support soon. However, this will just detect people. Doing the “identify” part is very hard. Like, you need a high resolution picture of the person’s face or their body shape needs to be unique.

Is there any easy hack you can use?