Number of Templates
During the enrollment process, 4 iris scan templates are created.
Creation of Template
After image capture of the iris, an algorithm quickly
defines the suitable area of the iris for feature
extraction and analysis. In this process, areas covered by
eyelids, deep shadows, and specular reflections are excluded.
The algorithm then automatically adjusts the width of the
pupillary boundary-to-limbus zone to maximize the amount of
iris analyzed, given varying ratios of pupil to iris sizes
due to different lighting in the environment. Then the
iris pattern goes through a calculation called
demodulation, in which complex-valued 2D Gabor wavelets are
used to extract the structure of the iris as a sequence of
phasors, whose phase angles are quantized to set the bits
in an IrisCode. It is not clear how Iridian technology
merge the 4 templates, but it is suspected that it would be
a similar process as fingerprints.
Below on the left is a
diagram that outlines the usable zone in an iris image. Below
on the right is a diagram illustrating the anatomy of the
eye, and location of the iris. Below in the centre is a
picture of a template with Iriscode.
  
Authentication Process
During the authentication process, when an iris is presented to the
image capture camera, an iris image is transformed into an
IrisCode in the same way the template is created. Any given
IrisCode is statistically guaranteed to pass a test of
independence against any IrisCode computed from a different
eye; but it will uniquely fail this same test against the eye
from which it was computed. Thus the key to iris recognition
is the failure of a test of statistical independence. After
this process, if the test fails, then the user is granted
access to the computer.
Several problems were encountered during enrollment.
Eye Alignment with the Camera and availability of Feedback
From the tester's experience, to align the eye with
the camera requires some practice. When the eye is at a good
distance and in focus with the lens, the light inside the
image capture lens changes from orange to green. When
the image is captured, the user can see three red lights
blinking at the bottom of the camera, and hear an audio
feedback that mimics a real camera when taking a picture.
All of this feedback provides users with some help when placing their eye in the correct
position, but it is a trail and
error process. The user has to position their about an arm's length
from the camera, and then gradually move toward the camera
until (hopefully) the light turns to green. There's no
guidance during the alignment process, only feedback when the
alignment is successful.
What makes alignment even more challenging is the fact
that there is no ideal location to place the camera. Though
the user guide suggested putting the camera on the monitor,
this does not produce an ideal height difference between the
user's eye and the camera. Note that the user's eye should be
just a little lower than the image capture lens. One
improvement might be to design a camera holder to mount the
camera on the monitor
Users with Glasses
The reflection of camera
illumination onto glasses surface, and the glare on the
metallic rim of the glasses caused by the lighting condition
in the room, both blocks the iris image and prevents successful
enrollment. Although some practice of positioning with the
camera can reduce the reflection, it's easier to just take off
the glasses. Without glasses, some users may not be able to
focus on the lens, but that is not a large concern. As long
as the lens can focus and is at a right distance to the eye,
the capture will be successful.
Squinty Eyes
Although the algorithm is claimed to
accommodate eyelids that partially covers the iris, in
practice the program rejects the image when the complete iris
is not visible. This is a problem for users who have small
and squinty eyes, such as the tester. Therefore, the tester had to
hold up her eyelids during the enrollment process, and it was
not a very comfortable experience.
Visibility of Feedback
| The image shown on the right is a capture of the
enrollment window showing live video feedback. This
feedback would be a great help to the user when adjusting
their position and to detect any problems such as glare and
reflection, if only the user could see it with their own
eyes while looking into up to the camera.
|
|
|
The photograph on the right was taken during product
testing. Notice how the user is unable to see the image
of their eye on the screen while focusing on the camera.
|
|
|
