The routine inclusion of a fingerprint scanner on certain popular brands of smartphone points to the fact that biometric security techniques are becoming an increasingly common aspect of our daily lives.
Unique personal identifiers based on physical characteristics or behavior are in many ways a more reliable – and certainly a more convenient – method of authentication and validation than traditional passwords, access codes, or security tokens. And iris patterns are one of the most distinctive and distinguishing personal attributes used in biometric security. In this article, we take a closer look at Biometrics and the use of Iris Pattern Recognition in cybersecurity.
Iris Pattern Recognition – The Eye Has It
The iris is a colored ring of muscle which opens and closes the eyeball of your eye like the shutter on a camera. The color of this ring-shaped region around the pupil is determined by the amount of the substance called melanin, which it contains. A higher amount of melanin in the iris produces a browner pigment, in a color spectrum that typically ranges from blue, gray, or greenish, to various shades of brown.
The pattern of the iris is established genetically in the womb but typically doesn’t manifest fully until the age of two. These patterns are complex, intricate, and extremely unique. The patterns of each of your two eyes are different from each other – and from those of anyone else. Even identical twins don’t have identical iris patterns.
This makes iris patterns an even more accurate biometric identifier than fingerprints (which are also unique to each finger, and each individual). But with wear and tear over time or due to injury, fingerprints undergo changes that may render them unreliable. It’s also been demonstrated that even replica fingerprints molded from gelatin are capable of fooling certain kinds of recognition technology.
Basics of Iris Pattern Recognition
To perform iris pattern recognition, one or more detailed images of the eye must be captured using a high-resolution digital camera. Images may be sampled at both the visible wavelength of light and the infra-red or IR end of the spectrum. Infra-red images are the type commonly captured in night photography, and it’s at this wavelength that night-vision apparatus like goggles and lenses operate. It’s also the wavelength where it’s possible to discern finer details in the pattern of brown and darker colored eyes.
A computer program known as a matching engine (which is configured with a special mathematical recognition formula or algorithm) is then used to compare the captured image of a person’s iris pattern to its own database of stored images. These matching engines can typically compare millions of images per second, with a level of accuracy comparable to conventional fingerprinting or digital finger scanning.
When used in access control or identity verification, iris pattern recognition requires the unique pattern of a subject’s iris to be positively matched against the system’s recognition database. This requires a two-stage process to be set up, in establishing the recognition ecosystem.
The enrollment phase is all about constructing a database of the iris patterns of all the people which the system will be required to validate or recognize. Each person must be photographed under visible and infra-red lighting conditions, to yield a pair of digital photographs for further processing and analysis.
The system removes unnecessary details (e.g., eyelashes) from each image, then catalogs around 240 unique identifiers. This is approximately five times as many points of comparison as the average fingerprint analysis yields.
These identifying features are then transformed into a simple, 512-digit number known as an IrisCode®, which is stored in the system database along with the person’s name and relevant credentials. The entire enrollment process typically takes a couple of minutes.
Verification occurs each time an individual has their eye photographed by an iris scanner, and their IrisCode® is successfully matched against the value stored in the system database. Note that this may not necessarily occur with the same scanner that was used for enrollment, depending on how the network is set up, how many geographical locations it covers, and so on. Iris patterns may also be compared using matching engines made up of hundreds, thousands, or even millions of individual records.
Limiting Factors of Iris Pattern Recognition
Much like the early technologies for facial recognition, proximity and image clarity are vital factors for iris recognition – both in capturing the initial enrollment photograph and in subsequent scans of a person’s iris pattern for matching against the verification database. Control mechanisms must also be put in place to guard against the substitution of high-definition photographs for actual irises at the point of scanning.
In ideal conditions, the subject should be within a few meters of the camera and must remain motionless or nearly motionless during the capture process. Additionally, any ambient lighting in the region of the scanner must not create reflections in the cornea (the shiny and transparent outer covering of the eye).
Certain kinds of contact lenses and spectacles may also produce distortions in the observed pattern of the iris.
These limitations emphasize the need for high-quality and high-definition digital imaging equipment and software. The software, in particular, has a role to play in correcting for minor errors such as motion blur, and in correctly demarcating the two boundaries (inner and outer) of the iris region. Systems also need to have the capacity to allow for changes in the surface area of the iris, as the pupil of the eye shrinks or grows in response to different lighting conditions.
Current Applications of Iris Pattern Recognition
Iris pattern recognition systems have been deployed at airports, border crossings, and individual points of entry or exit for buildings in several countries across the globe. Devices may be wall-mounted (for use in airports and other buildings) or hand-held and portable (as in the iris scanners deployed by the US Army).
Iris recognition systems for personal use are available for protecting laptops and other equipment, while a number of mobile apps are available for providing access control and anti-theft protection on smartphones and other devices fitted with front-facing cameras.
A small portable iris-scanning device is available on the consumer market, for personal applications such as logging onto secure websites without having to use a password.
The Vision Ahead
It’s been estimated that iris pattern recognition is ten times more accurate than fingerprinting. And since iris patterns remain relatively unchanged for decades, this recognition accuracy has the potential to stretch for many years into the future.
Despite the somewhat high initial costs and relative newness of the technology, it’s likely that it will continue to evolve to the stage where photography and digital image processing may be conducted with more clarity and at greater distances. This kind of evolution has the potential to raise issues over privacy concerns, civil liberties, and the ways in which recognition data is used and handled.
These issues aside, in terms of its accuracy, uniqueness in identifying an individual, and overall convenience over more conventional methods of verification and access control, iris pattern recognition has much to offer, in cybersecurity applications.
Share this Post