Traditional authentication systems require the user to perform the cumbersome task of memorizing numerous passwords, personal identification numbers (PIN), pass-phrase, and answers to secret questions. Biometric authentication comes in play to release the users from the difficulties of remembering and protecting passwords.
To date, different biometrics have been researched and used, such as fingerprint, hand geometry, face, odor, DNA voice, ear, gait, etc.
Among all the biometrics in use today, eye biometrics (iris and retina) offers the highest level of uniqueness, universality, permanence, and accuracy, since iris and retina are an individual’s unique biological traits. Iris Recognition and Retina Scanning are the physiological biometrics methods for identifying a person or verifying a person’s identity.
Iris is the visible ring on the front side of the eye that surrounds the pupil of one’s eye. Each iris is unique, and even irises of identical twins are different. Not only the iris patterns are unique for each individual, but irises of left and right eyes of the same individual are also unique. This uniqueness holds in family siblings, and even identical twins, where other genetic details such as facial appearance are so similar. The iris’s complex structure carries distinctive information, and under normal health conditions, it remains unchanged from early childhood to death of the individual. This stable uniqueness of iris texture becomes the basis of iris-based biometric recognition.
Ophthalmologists Leonard Flom and Arin Safir first noted the distinctive features of iris and described methods for iris recognition in 1987. Dr. John Daugman of Cambridge University later developed the algorithms, mathematical methods, and techniques to encode iris patterns and efficiently compare them. All commercial applications currently implement Daugman’s patented techniques. Daugman’s approach uses 256-byte IrisCode as a template and performs matching applying XOR operations on the IrisCodes.
The major phases in the iris pattern recognition process include iris scan for image acquisition, iris localization for distinguishing the iris from the rest of the eye, feature extraction, and template creation, followed by matching. The very high accuracy, attained in iris recognition today, suggests it to be a very promising biometric for authentication purposes.
- High accuracy
- Iris does not change over time
- The method does not require intimate contact with the reader
- Iris patterns are more robust than voices.
- A higher average for matching performance
- Convenient for people who wear glasses
- Low chances of a false positive
- Almost unaffected by the environment due to being protected by the cornea and the aqueous humor.
- Acquiring an image requires proper alignment and positioning.
- The result might be affected due to pupil size change.
- Hard to use
- Difficult to integrate with other systems
- The position of the eye could be problematic.
- Require specialized devices so that it can be expensive.
The retina is the innermost layer of a human eye. This layer on the back of the eye has complex blood vessels and nerve cells, creating a pattern beneath the surface that is unique to each individual. This blood vessel pattern is a reliable biometric characteristic for retinal scanners.
The first study to confirm the uniqueness of the retina’s blood vessel pattern was released in 1935 by Dr. Carleton Simon and Dr. Isodore Goldstein. They laid out their discovery that every retina possesses a unique blood vessel pattern. They later released their paper, suggesting the use of photographs of these blood vessel patterns of the retina to identify people. The second study by Dr. Paul Tower came out in the 1950s. He discovered that even among identical twins, the blood vessel patterns of the retina are unique. In 1981, the first real prototype of a retinal scanning device was developed. It used infrared light to illuminate the retina’s blood vessel pattern.
The process of enrollment and verification/identification in a retinal scanning system is the same as the process for other biometric technologies. The process involves the acquisition and processing of images, unique feature extraction, and template creation and matching. Experts acquire digital images of retinal patterns by projecting a low-intensity beam of visible or infrared light into a person’s eye and scanning the retina.
- Very accurate
- Impossible to forge a retina
- Has lower error rate (1 in 10,000,000) as opposed to fingerprint identification (1 in 500)
- Low false rejection rate and low false acceptance rate
- Inconvenient for people who wear eyeglasses
- Uncomfortable for some users
- The method is fairly new, and not many use retina biometric devices.
- Both iris and retinal pattern recognition are eye-based biometrics,
- They need the users’ cooperation for image acquisition executed by putting the open eye in front of a digital camera.
- Both techniques use near-infrared light to illuminate the object (iris or retina) of interest.
- Patterns in the iris and retina are unique not only for each individual, remains intact throughout life under normal health condition.
- Since they are non-contact biometrics, both techniques are less vulnerable to identity theft than fingerprint, voice, etc.
- They are pattern-dependent, not sight-dependent. Even blind people can enroll successfully.
- Iris image capturing uses a video camera, while retinal scanning devices require specialized cameras used in ophthalmology. Consequently, retinal scanning devices are costly.
- Iris recognition obtains an image of the external eye, i.e., the iris surrounding the pupil, the white portion of the eye, the eyelids, and possibly the eyelash as well. In contrast, retinal pattern recognition obtains the internal part of the eye, i.e., the retina’s vascular pattern.
- The intricate iris patterns carry an astonishing amount of information, including 200 unique spots, while retinal scanning gathers unique features from the blood vessels.
- Iris pattern recognition tends to be more acceptable to users compared to retinal pattern recognition.
- Since the retina is an internal protected organ, it requires specialized ophthalmologic cameras for retinal image acquisition, making it impossible to spoof the retina to date. But iris spoof attacks are possible even with a high-quality photograph. Using a contact lens is another method to spoof some iris sensors successfully.