LOL. this is a funny thread. I don't know how you can trick the Kinect into logging in the same person under multiple accounts. Maybe you shave your head for one, wear a beanie for another, wear a beard for another?
for people having real facial recognition problems:
What Kinect ID really is is a database of images of the person where the system compares the person to the database of images to make the identification. So the more pictures it takes at varying lighting conditions and angles, the more accurate it is. Meaning you need to run the setup multiple times, at different times of day (night time lighting is obviously is different than 8am is different than high noon), different lighting conditions, different angles (Kinect at low height, Kinect at high height).
I've run the Kinect ID program with the Kinect at 6 feet (above the TV) and at 4 feet (below the TV). With all the lights of the family room on, with half the lights of the family room on, with bedroom lights on, etc. At 9AM, noon, 5PM, 8PM.
And the database is self-contained on each individual 360. So if you move the Kinect between your bedroom 360 and your family room 360, you have to build the databases separately on each 360. Obviously, the height of the Kinect in the bedroom will differ from the height in the family room (different size TVs, different entertainment centers, etc).
It's not necessary to do all this (and honestly is a bit excessive) but if you want the most accuracy, you do it. Also, if you wear glasses sometimes but contacts other times, run the Kinect ID with both your glasses on and off. The more images the database has, the more accurate it is. I've even run the KinectID with my wireless headset on and with it off.
However, we don't know the limit so don't go crazy and take a zillion pictures because the newer pictures you take may overwrite the old pictures. Take the pictures that will be most common for that person.
If you've ever used photo recognition software like Picasa or Windows Live Photo Gallery, you know that the more pictures it takes, the more accurate it is at identification.