How to ensure user has lifted up the phone to eye-level?

We are creating an app wherein we need to detect that the person has lifted the phone up to eye-level. Something like this photo below:

We were thinking to use the camera to detect the face (as in a selfie) however, a person can still take a photo of his face from below (without lifting his arms to this level) so it's not so optimal.

I wonder what methods we can use or what instruction to give to ensure that this has been done by the user.