I can suggest using QR codes or something like that, that the person must use and put in some place visible, so a camera can parse the QR code (instead of "parsing" a face) and identify the person. I think there is no much waste of time, just a second for the person to stand still so the camera can parse the QR code.