Retinal recognition – The ultimate biometric
Review of the last article
As we have seen, there are numerous applications for Biometric Technology. They can range from very simple applications, such as Single Sign-On Solutions, all the way to the most complex, such as that of the e-Passport infrastructure.
As it was discussed, the e-Passport is quickly becoming the replacement for the traditional paper passport. The primary reasons for this are for both security and convenience. Regarding the former, the paper passport can be, to a certain degree, easily forged, replicated, or even hijacked. With the latter, processing a paper passport can take time, especially at the customs and immigrations line at the country of destination.
What should you learn next?
What should you learn next?
In contrast, the e-Passport is a much secure identity document. For example, it contains a microchip which can store up to three Biometric Templates of different modalities, thus offering a true multi-layered approach to security. It is also very convenient because it consists of an RFID antennae embedded into the microchip. Because of this, a traveler's identity can be instantaneously confirmed in just a matter of seconds, by merely "flashing" the e-Passport in front of the e-Passport reader.
But despite these key, strategic advantages, the e-Passport infrastructure is also prone to Cyber based attacks, given its heavy dependency upon a networked environment. Thus, the need for Penetration Testing across all levels is required to make sure that all of the hidden security vulnerabilities can be discovered and quickly patched up.
One of the Biometric Templates which is used quite heavily in the e-Passport infrastructure is that of Iris Recognition. Another similar Biometric modality is Retinal Recognition, which is the topic of this article.
The anatomy of the retina
Throughout the entire human anatomy, it is the eye which captures the intrigue of both medical practitioners and research scientists. As a result, it has also received great interest as well from the Biometrics Industry as a means to positively confirm the identity of an individual.
The Iris is located in the front of the eye, and its primary function is to allow the requisite amount of light into the eye which is necessary to maintain a proper line of vision. The Retina is the grouping of the blood vessels in the back of the eye, which connects to the front of the optic nerve which then goes into the brain.
The primary function of the Retina is to feed visual information and cues into the brain so that it can further process.
In other words, the retina "is to the eye as film is to a camera". (SOURCE: 1). On a much more scientific level, the Retina possesses multiple layers of sensory tissue as well as millions upon millions of photoreceptors. Their primary function is to transform the light rays captured by the Iris into various types of electrical impulses from which they are then converted over into decipherable images.
The Retina possesses two types of photoreceptors, which are: 1) the Rods; and 2) The Cones. It is the former which allows for both night and peripheral vision to take place.
An overview into retinal recognition
The interest of using the Retina as means for verification and/or identification goes back all the way to the 1930s. The first scientific study was conducted by Dr. Carleton Simon and Dr. Isodore Goldstein. In their work, they describe how every individual possesses a unique Retinal structure, because of the differing distribution of blood vessels.
This was the first study actually to suggest the use of the Retina as means to help confirm the identity of an individual. In fact, their published scientific paper, which came out in 1935, is considered to the be the formal birth of Retinal Recognition.
The next scientific study which confirmed the uniqueness of the Retina was conducted in the 1950s by a scientist known as Dr. Paul Tower. It was discovered through his work that even identical twins have a very distinct and unique set of Retinas.
Not only because of the uniqueness it possesses, the Retina also contains very rich information about an individual, just as much as the DNA Strand. Also like the Iris, the Retina hardly every changes over the lifetime of an individual unless they are afflicted with a disease or blindness.
Thus, it is considered to be a very stable Biometric modality, when compared to some of the others (such as Facial Recognition or even Fingerprint Recognition). Because of these strategic advantages, Retinal Recognition has been termed as the "Ultimate Biometric of All".
The process of retinal recognition
Unlike the other Biometric modalities, Retinal Recognition requires a great amount of cooperation from the end user to collect high quality, raw images. This requires that the end user must be in very close proximity to the Retinal scanning device. This is in sharp contrast to Iris Recognition, where images can be collected from a very far distance.
The process can be broken down as follows:
- Image/Signal Acquisition and Processing:
In this first phase, the individual must place their eye into a tiny receptacle. From here, an infrared light beam light is then "shot" into the eye, in order to fully illuminate the Retina. To get the best quality, raw images of the Retina, this light is shone into the eye in a 360-degree manner. This process can take as long as a few minutes. In order to make sure that resilient images are captured, the individual must remain perfectly still while the infrared light completes its rotation. Any eyeglasses must be removed, in order to avoid any signal interference (in other words, this would cause the infrared light beam to reflect back into the scanning device without having captured any actual images). At this stage, up to five raw images can be collected and analyzed, in order to create the composite image from which the unique features will be extracted.
- Enrollment and Verification Template Creation:
In this second phase, the unique features are then extracted. Genetic factors actually do not dictate the composition of the blood vessel pattern which make up the Retina. In other words, this is not written into the DNA structure of the individual and is not handed down to the offspring. Because of this, up to 400 unique data points can be captured from the Retina, unlike the fingerprint, where only 30-40 unique data points can be captured. After this has been done, the Enrollment Template is then created. The size of the Retinal Enrollment Template is only at 96 bytes. Thus, it is also deemed to be the smallest Biometric Template of all. This is, of course, has numerous advantages to it. First, there is much lower processing overhead required when examining the statistical closeness between the Verification and Enrollment Templates. Second, this small size means that a greater number of templates can be stored in a single database, and optimizing it will be much more efficient when compared to a database of Facial Recognition Templates, where the sizes are much larger. This same process is also used to create the Verification Template.
Because Retinal Recognition requires such a close proximity to the end user, there are a number of variables or factors which can greatly influence the quality of the raw images which are collected. This, in turn, can also have a huge impact on the quality of the Enrollment and Verification Templates as well.
These factors are as follows:
- Lack of cooperation on part of the individual:
As it has been described, the individual must remain perfectly still throughout the above-described process, especially when the raw images are being captured. Any sudden or unintentional movement can negatively affect the lens alignment of the aperture which is used to transmit the infrared light beam into the Retina.
- The distance between the individual's eye and the lens is too far:
For a high-quality scan to take place, there must be a distance of not greater than 3 inches between the receptacle and the Retina. If the distance is greater than this, then the scanning process must be repeated over again until it is correct. In this regard, when compared to the other Biometric modalities, precision is of utmost importance in Retinal Recognition.
- The size of the individual's pupil:
A smaller than average size pupil can greatly reduce the amount of external light which is transmitted to the Retina. This problem can become further exacerbated if the pupil constricts even further because of inconsistent lighting conditions.
Retinal recognition: The advantages and disadvantages
Just like the other Biometric modalities, Retinal Recognition has its own set of advantages and disadvantages. They can be described as follows:
- The Retina is deemed to be very stable, and hardly ever changes over the lifetime an individual. Thus, in this regard, it is considered to be the most reliable Biometric Technology which is available in the marketplace today.
- Given the small file size of the Retinal Recognition Templates, the time it takes for the system to confirm the identity of an individual is very quick; it can take place in just less than two seconds.
- Because of the high number of unique data points the Retina possesses, there is almost no error that when an identity of an individual is confirmed, it is truly that person. In other words, the statistical probability of an impostor being falsely accepted by a Retinal Recognition System is almost nonexistent.
- Since the Retina is located from within the structure of the eye itself, it is not prone to the harshness of the external environment like Hand Geometry Recognition and Fingerprint Recognition.
- Overall, there is a very negative attitude about using Retinal Recognition amongst the public, at least here in the United States. For example, because of the sheer intrusiveness which is involved with it, many people perceive that it poses a serious health risk to the eye. There have been no documented cases in this regard.
- There is a very strong unease about having to place the eye into a receptacle and having an infrared light beam being shone directly onto it.
- When compared to all of the other Biometric modalities, Retinal Recognition demands the highest levels of cooperation and motivation from the end user to capture high quality, raw images. As a result, the Ability to Verify metric can be as low as 85% (other modalities are as high as 99% or even 100%),
- Because of the attention that is required by the end user, it can take numerous attempts and a long time to get the results which are required. Thus, as a result, if the process is not done correctly, it can lead to a very large False Rejection Rate (this occurs when a legitimate individual is improperly denied access to either physical or logical based resources by the Retinal Recognition System).
Overall, the effectiveness and viability of Retinal Recognition can be examined against seven criteria which are also used by the Biometrics Industry here in the United States:
Each and every individual, unless they are afflicted with blindness or some other serious disease of the eye, has a Retina, so thus, it can be scanned.
Except for the DNA Strand, it is the Retina which possesses the most number of differing data points in the entire human anatomy.
Unless an individual has diabetes, glaucoma, high blood pressure, or cardiac disease, the Retina will hardly every change regarding structure or physiological makeup during the lifetime of the person.
Even if the individual is fully cooperative, it can still be difficult to a collect a high quality, raw image of the Retina. This is because the scan area is so small, when compared to the other Biometric modalities.
Because of its stability, the Retina possesses extremely high levels of accuracy. In fact, under optimal conditions, the error rate can be as low as 1 in 1 million.
As mentioned earlier, the acceptance of Retinal Recognition by the general public is extremely low.
- Resistance to circumvention:
Because of its stability and richness, it is almost impossible to spoof a Retinal Recognition system.
Taking all of this into consideration, the market applications are extremely limited for Retinal Recognition. Thus, Retinal Recognition is used primarily in Physical Access Entry applications, where extremely high levels of security are required. This includes military installations and bases, nuclear facilities, and laboratories where very high caliber research and development is taking place.
Government-based security applications also, to a certain degree, also make use of Retinal Recognition. One of the best examples of this is the State of Illinois, where this modality was used in identifying welfare recipients to reduce fraud (this would occur when an individual would use several aliases to receive multiple payments).
What should you learn next?
What should you learn next?
Overall, it is expected that the use of Iris Recognition will grow rapidly over a short period of time, but Retinal Recognition will still only occupy that very limited of pool of applications just described, and not grow any further from that point.
Biometrics Personal Identification in Networked Society: Anil Jain. Ruud Bolle, and Sharath Pankati.