Iris recognition is an innovative and secure biometric authentication method. Artificial intelligence makes this technology more accessible for use in CCTV cameras, smartphones, and other access and security controls. Such identification reduces the risk of failure of facial recognition systems. In this article, we’ll cover how the technology works, a comparison of the iris and retina scanning, and the prospects for the future of iris recognition.
Iris authentication technology is believed to have evolved from another very well-known technology, retinal authentication. Iris scanning technology was first proposed in 1936 by an ophthalmologist, Frank Burch. He stated that each person’s iris is unique. The probability of its coincidence is about 1078, which is much higher than with fingerprinting. According to the theory of probability, in the entire history of mankind, there have not yet been two people with the same iris. In the early 90s, John Duffman of Iridian Technologies patented an algorithm to detect the iris of the eye.
Scientists have conducted several studies showing that the human retina can change over time while the iris remains unchanged. It is impossible to find two completely identical patterns of the eye’s iris, even in twins.
Glasses and contact lenses, even colored ones, will not affect the imaging process in any way. It should also be noted that the performed operations on the eyes, removal of cataracts, or implantation of corneal implants do not change the iris’s characteristics; it cannot be altered or modified. A blind person can also be identified using the iris of their eye. As long as the eye has an iris, its host can be identified.
The camera can be installed at a distance of 10 cm to 1 meter, depending on the scanning equipment. The term «scanning» can be misleading, as the process of obtaining an image is not scanning but simple photographing.
The iris’ texture resembles a network with many surrounding circles and patterns that can be measured by a computer. Iris scanning software uses about 260 anchor points to create a sample. In comparison, the best fingerprint identification systems use 60-70 points.
The cost has always been the most significant deterrent to technology adoption, but iris identification systems are becoming more affordable for various companies. Proponents of the technology claim that iris recognition will very soon become the mainstream identification technology in various fields.
The unique iris pattern must be recognized to pass such a biometric scan, allowing an identification response. This means that there are two stages in scanning the iris of the eye:
All that such a recognition system needs to verify a person is a snapshot of his iris. Therefore, to launch this scenario, each person must go through one-time photographing of their eyes. In this case, the photographing procedure takes place under standard lighting and invisible infrared – the latter is a type of light used in night vision devices that has a slightly longer wavelength than ordinary red light.
Infrared light in the iris scan helps to recognize the unique pattern of darker eyes more accurately, which is more challenging to do in normal light.
These two digital photographs, taken under different types of lighting, are then subjected to computer analysis, which removes unnecessary details (such as eyelashes) and highlights about 240 features in iris patterns (approximately five times more «features to compare» than used in fingerprint identification systems).
After that, all found unique characteristics for each eye are converted into a simple, digital number consisting of 512 digits (also called an iris code), which are stored in the computer database along with your name and other details. Taking a picture of your eye is completely automatic and does not take more than a few minutes.
Once the iris photo has been entered into the database, your identification process will be easy and hassle-free. The person stands in front of an iris scanner connected to the database and goes through a quick re-photographing of the eye.
The system quickly analyzes the resulting image, extracting an iris code from it. It starts the procedure for comparing hundreds, thousands, or even millions of iris codes available in the database. If the code matches one of the codes entered in the database, the person is positively identified. If it fails, this means that the snapshot is not familiar to the system, or the person is merely trying to counterfeit who they are.
The iris is a circular piece of colored muscle tissue that frames the human pupil and helps it to clench/unclench like a camera shutter. Our iris’ color pattern is formed at the genetic level even when we are in the womb, but it finally completes its formation during the first two years of our life. The color of the eye’s iris depends on the amount of melanin pigment: the more melanin, the more the eyes have a brown tint, the less — the more pronounced the blue color.
Although we are used to distinguishing the color of each person’s eyes clearly — «brown eyes,» «green eyes,» «blue eyes» — the color and pattern for each particular iris is unique. For example, even two people’s eyes have two different shades and patterns of their irises. The same goes for the eyes of genetic twins.
|HD Criterion||Odds of False Match|
|0.26||1 in 1013|
|0.27||1 in 1012|
|0.28||1 in 1011|
|0.29||1 in 13 billion|
|0.30||1 in 1.5 billion|
|0.31||1 in 185 million|
|0.32||1 in 26 million|
|0.33||1 in 4 million|
|0.34||1 in 690,000|
|0.35||1 in 133,000|
Automatically detecting new types of cosmetic contact lenses in iris images is a highly complex pattern recognition task. But recently, experimental datasets have emerged to help researchers investigate the problem. Given the pace of progress in other aspects of iris recognition, the research community is likely to make rapid progress in addressing the tampering problem. You can also use retina recognition to improve recognition accuracy.
The iris is one of the unique biometric characteristics used for identification. During verification, about 260 key points are used (in comparison, fingerprint verification uses about 16 key points). Simultaneously, the template itself takes up a small amount of memory, which allows you to quickly authenticate a user and use massive databases with relatively little computing resources.
Access control and accounting systems with iris identification have FAR — 0.00001% and FRR — 0.016%. According to the NIST (National Institute of Standards & Technology), iris recognition accuracy is 90-99%. ScienceDirect has also conducted a study that showed 100% effectiveness using the iris recognition method.
It is believed that it is impossible to forge identification data using this method. The fact is that, in addition to the individual pattern of the iris, the human eye has unique reflective characteristics (due to the state of tissues and natural moisture), which are considered in the process of reading information.
And to further enhance safety, some iris scanners also capture the eyeball’s involuntary movements inherent in a living person. By the way, authentication by the iris of a dead person is also considered impossible: after death, the pupil expands, making the iris area too narrow and therefore unsuitable for scanning.
This biometric characteristic is unlikely to change over time: the only reasons could be medical surgery or severe injury.
There are concerns regarding iris identification, fearing that an iris scan’s infrared rays could negatively affect vision. Our eyes do not have protective reactions to infrared radiation. When rays of bright light blind us, we reflexively squint or turn away, and the pupil of the eye narrows spontaneously. Since we do not see infrared light, we cannot determine when we fall under its influence, and the eyes do not respond to this radiation by constricting the pupil.
To reduce the harmful effects of infrared light on the eyes, designers use visible white light before infrared scanning. The use of such illumination causes the pupil to contract spontaneously, which reduces the penetration of infrared rays into the cornea of the eye. Another positive aspect of the pupil’s constriction when identifying by the iris of the eye is the expansion of the identifiable area. The increase in the iris’ visible area allows you to get unique information for its encoding and recording in the biometric template.
Conventional photo and video cameras of telephones and cameras have a built-in IR-cut filter designed to exclude infrared radiation’s influence on the quality of the resulting image. Biometric facial identification from the front-facing 2D camera is easy enough to deceive. To detect deception, the developers began to use point IR illumination, with the help of forming a depth map of the object being shot.
Controlling a three-dimensional figure in front of the camera prevents simple methods of deceiving biometric identification systems using a photograph or video recording of an identified person.
This happens due to the lack of an IR filter in the front cameras of most modern smartphones.
Manufacturers are implementing solutions that minimize the harmful effects of infrared radiation on the eyes:
The human retina is the thinnest tissue in the body and is made up of nerve cells located at the back of the eye. Due to the complex arrangement of the capillaries that feed the retina with blood, each person’s retina is unique. The network of blood vessels in the retina is so complex that it differs even in identical twins. The retinal pattern may change due to diseases such as diabetes, Mellitus, or glaucoma. However, in other cases, the retina, as a rule, remains unchanged from the moment of birth until death.
Retinal scan technology is used to display a unique pattern of human retina. The retina’s blood vessels absorb light more intensely than the surrounding tissue, so they are easy to identify. Retinal scanning is carried out by projecting an invisible beam of infrared light into the human eye through the scanner’s eyepiece.
Since the retina’s blood vessels absorb this light more intensely than the rest of the eye, a pattern is created during the scan, which is converted into computer code and stored in a database. Retinal scans also have medical applications. Infectious diseases such as AIDS, syphilis, malaria, and chickenpox, as well as hereditary diseases such as leukemia, lymphoma, and sickle cell disease, affect the eyes. Pregnancy also affects the eyes. In addition, signs of chronic diseases such as chronic heart failure or atherosclerosis also appear first in the eyes.
Often confused with retinal scanning, iris recognition systems capture the eye’s image and then analyze the colored portion around the pupil, the iris that you can see with the naked eye.
In turn, the retina is made up of photoreceptor cells located at the back of the eye and cannot be seen. While iris recognition captures the iris’ texture pattern, a retina scanner captures an image of the network of blood vessels inside the eye.
|Disease||Safe||Low probability of causing harm|
|Distance||Normal distance||Close distance|
|Complexity||Easy to use||Complex|
|Acceptance||Highly accepted||Lowly accepted|
«Unlike the retina, the iris can be seen with the naked eye, so it is much easier to get a high-quality image of the iris,» says David Usher, a senior research scientist at Retica Systems, which designs and develops identity data systems based on iris analysis. «In the iris recognition system, images of the iris are captured using near-infrared (NIR) LEDs and algorithms, which are then used to convert the texture of the retina into a special code. This code or image is compared with patterns in the device’s memory, and then confirmed or refuted personal identification.»
Iris recognition is more widely accepted as a commercial biometric technique than retinal scanning. While both recognition methods are non-contact, retinal scans are considered invasive because they direct visible light into the eye, whereas iris recognition uses contactless digital photography for identification.
The iris identification technology is more widely used in various industries. To date, nearly 1 billion people worldwide have registered with iris recognition systems for security and convenience purposes such as national identification, border control, finance, banking, etc.
The iris recognition by-product market is divided into smartphones, tablets, PCs/laptops, and scanners. The widespread adoption of eye scanners in banking and finance, military and defense, and travel and immigration, especially for identity management and access control applications, is the main reason for the largest share of this segment in the iris recognition market.
Recognition of people by the pattern of the iris of the eye nowadays finds more and more widespread use. In some respects, it still lags behind other biometric technologies; in others, it overtakes them. Simultaneously, biometric technologies have many common advantages over other methods of human identification.
Biometric technologies can be used as an alternative to existing authentication methods that require memorizing countless passwords, passphrases, PIN-codes of plastic cards, bank accounts, and so on.
Today, the use of such technologies is most often performed in security systems for:
Biometric data read from the iris plays an important role in controlling access to highly restricted areas. Twenty-nine airports in Canada are using iris recognition technologies to verify that they are authorized to board an aircraft. At Amsterdam Airport Schiphol, a fast-track passport control system uses iris scanning to identify the crew and passengers who fly frequently.
Most installations of iris recognition technology at airports are used for passengers on international flights and can be used instead of showing their passports. Recognition is usually accomplished in less than one second, and the high resolution and quality of the resulting images reduce the rate of rejection and rejection errors.
One of Boston’s largest hotels uses an iris-based access control system to identify guests staying in the elite presidential suite. Another facility in Boston uses this technology to keep track of children who go missing and need to be identified in the future.
Many governments have already used this technology as a highly useful, secure identity management tool. It is soon expected to enter mainstream commercial sectors such as automotive and mobile communications.
As a result, the iris pattern’s application is expanding, although its potential is just beginning to unfold. Each of the leading manufacturers offers scanners in several form factors and can equip large corporations and security agencies with the flexibility to accommodate all their needs.
Let’s underline the most prominent features of the technology.
Iris recognition is the most advanced technology available today. What are its advantages?
It is worth noting that iris recognition has some disadvantages:
Iris recognition remains one of the most promising biometric technologies for personal recognition. Especially in demand is the realization of the iris’ potential for use in non-contact scenarios identification and a face image — and possibly other contactless biometric identifiers.
Therefore, the most relevant research directions are to improve recognition in non-invasive scenarios due to improving sensors, improving the system’s informative signs, and integration with other modalities. Of particular interest is the use of the iris in cryptographic applications and secure identification.
Eye scan recognition has certain advantages over other biometric technologies that make this technology one of the most preferred mobile devices. In recent years, several companies have introduced smartphones equipped with iris authentication technology. Biometric authentication is a promising technology that will eliminate the usual authentication schemes using a password. This will increase the convenience of working with the device, and at the same time, increase the level of protection of personal data.
Iris recognition remains one of the most promising biometric technologies for personal recognition. Especially in demand is the realization of the potential of the iris for use in non-contact scenarios identification together with a face image and possibly other contactless biometric identifiers. Therefore, the most relevant research directions are to improve recognition in non-invasive scenarios due to improving sensors, improving the system’s informative signs, as well as through integration with other modalities.
When scanning an eye, the pupil region and the iris itself are highlighted. The resulting ring is programmatically cleared of noise and converted into a rectangular format — an iris code, containing information about the object’s unique characteristics in black and white (like a barcode or QR code). Then the iris code is compared with the database of registered templates. At the same time, the processing speed is too high, which makes it possible to use the system for working with large databases.
Iris recognition is the most advanced and accurate biometric technology available today, providing precise identification without PINs, passwords, or access cards. User registration is done in less than 2 minutes. Authentication lasts no more than 2 seconds.
Infrared radiation increases the overall temperature of the aqueous humor in the eye, affecting the cornea and aqueous humor. IR-A radiation is absorbed by the retina and is very ineffective in damaging the retina. Moreover, the image capture procedure is concise and can usually be achieved within 2-10 seconds. However, in today’s biometrics industry, security standards can be ignored.
Although the term “scan” is often used to refer to iris recognition, it is not a scan at all. The technology is based on recognition of the iris pattern, and the pattern capture technique, in turn, is based on video recording. The camcorder does not require bright lighting or close-ups.
Iris recognition technology is used to gain access to devices and databases and access buildings, rooms, and apartments. However, the most massive increase in installations of this technology is now observed at the airport. Various fields of application in iris recognition technologies and targets and objects of injections are presented in the table.
Colorless contact lenses are not believed to interfere with iris recognition systems. However, research by Sarah Baker, Amanda Hentz, and Kevin Bowyer of the University of Notre Dame, Indiana, has shown that this is not the case. Contact lenses create significant artifacts in the image and lead to significant degradation of the biometric solution’s accuracy.