Facial Scan in Biometrics

Image Acquisition

Iris-scan technology requires the acquisition of a high-resolution image of the eye, illuminated by an infrared imager, in order to effectively map the details of the iris. The acquisition process and the amount of effort required on the part of the user differ according to the type of acquisition device used. The three major types of iris-scan systems are kiosk-based systems, physical access devices using motorized cameras, and inexpensive desktop cameras. Although-iris-scan vendors do not emphasize their use of infrared light, each system does rely on infrared imaging using wavelengths in the 700-to 900-nm range (judged to be safe by the American Academy of Ophthalmology).

Kiosk systems

It requires that users stand approximately 2 to 3 feet from the camera or cameras, which are positioned at the height of a typical user’s eyes. The user must remain still, as it is difficult for these systems to locate irises on a moving target. When faced with a user in its field of view, the kiosk-based camera searches for eye shapes. In order to facilitate this, users may need to remove their eyeglasses, because glare from eyeglasses can impact the ability of the systems to match users. From this point, the acquisition is automatic – the system normally locates the iris and acquires the image within 1 to 2 seconds.

Physical access devices

It requires slightly more user effort. A small camera mounted behind a mirror acquires the image; the user locates his or her eye in the mirror, centering the iris within a 1-inch by 1-inch square. The user may also be vocally prompted to move slightly forward or backward to enable image capture. The proper distance from the mirror is approximately 3 inches. A high-quality camera focuses on the eye, acquiring a series of images until a resolution threshold is met. Acquisition through these devices is more challenging because it is contingent on users’ ability to follow interactive prompts. Also, individuals who favor a particular eye can have problems locating their weaker eye in the mirror.

Desktop cameras

It is used for logical access, and is the newest type of iris-scan device. Acquiring the image at a distance of approximately 18 inches the device requires the user to align his or her line of sight with a guidance light or hologram. When the user is positioned correctly, the camera acquires the image. These systems have proven fairly difficult for some users, who find it difficult to orient themselves at the proper distance from the camera.

Image Processing

Regardless of the acquisition device, the process of mapping the iris remains the same. After the camera locates the eye, an algorithm narrows in from the right and left of the eye to find the iris’s outer edge. The iris-scan algorithm then locates the inner edge of the iris at the pupil. Locating the iris-pupil border can be challenging for users with very dark eyes, as there may be very little difference in color as rendered in the technology’s 8-bit grayscale imaging.

Once the parameters of the iris have been defined, a black-and-white image of the iris is used for feature extraction. The core technology can account for pupil dilation, occlusion due to eyelids, and reflections due to the acquisition camera; the area used for feature extraction is a horizontal band extending from the far left to the far right of the iris. When the pupil dilates, the iris patterns shrink and expand in a normalized fashion such that algorithms can translate a dilated verification to a nondilated enrollment.

Distinctive Features

The patterns that constitute the visual component of the iris are surprisingly distinctive. A primary visible characteristic is known as the trabecular meshwork, a tissue that gives the appearance of dividing the iris in a radial fashion. Other visible characteristics include rings, furrows, freckles, and the corona. Iris patterns are formed before birth and remain stable throughout an individual’s lifetime (unless one suffers an eye injury). Tests have shown that individuals’ left and right eyes have different iris patterns, and that even identical twins’ irises have almost no statistical similarity.

Iris-scan algorithms map segments of the iris into hundreds of independent vectors. The characteristics derived from iris features are the orientation and spatial frequency of distinctive areas (the what) along with the position of these areas (the where). Not all of the iris is used; a portion of the top as well as 45 degrees of the bottom are unused to account for eyelid occlusion and reflections.

Template Creation

The vectors located by the iris-scan algorithm are used to form enrollment and match templates, which is generated in hexadecimal format as opposed to binary. Depending on the iris-scan solution, between one and four iris images may need to be captured for enrollment and template generation. The use of multiple images ensures that the data extracted to form a template is consistent and that there are no reflections being misinterprinted as iris features.

Template Matching

Iris-scan solutions generally perform identification as opposed to verification, meaning that the best match template is compared against all system enrollments to find the best match. Althoughidentification is much more challenges than verification, the process is normally very brief: On some processors, iris-scan technology is capable of searching hundreds of thousands of records per second. Because of the distinctive nature of the iris, identification deployments are not designed to return candidate lists but high-confidence matches.


Compiled By: Chaudhary Amit V.