Accuracy of Biometric Techniques

It is important to understand fully the accuracy of biometric devices before deploying it. Biometric like other technologies is not 100% accurate but it enhances security level and makes it difficult to breach the security.

The three key performance metrics are
1. False match rate (FMR)
2. False non match rate (FNMR)
3. Failure to enroll rate (FTER).

False Match Rate (FMR)
A false match occurs when a system incorrectly matches an identity, and FMR is the probability of individuals being wrongly matched. In verification and positive identification systems, unauthorized people can be granted access to facilities or resources as the result of incorrect matches. In a negative identification system, the result of a false match may be to deny access. For example, if a new applicant to a public benefits program is falsely matched with a person previously enrolled in that program under another identity, the applicant may be denied access to benefits.

False Non Match Rate (FNMR)
A false non match occurs when a system rejects a valid identity, and FNMR is the probability of valid individuals being wrongly not matched. In verification and positive identification systems, people can be denied access to some facility or resource as the result of a system’s failure to make a correct match. In negative identification systems, the result of a false non match may be that a person is granted access to resources to which she should be denied. For example, if a person who has enrolled in a public benefits program under another identity is not correctly matched, she will succeed in gaining fraudulent access to benefits.

Failure to Enroll Rate (FTER)
FTER is a biometric system’s third critical accuracy metric. FTER measures the probability that a person will be unable to enroll. Failure to enroll (FTE) may stem from an insufficiently distinctive biometric sample or from a system design that makes it difficult to provide consistent biometric data. The fingerprints of people who work extensively at manual labor are often too worn to be captured. A high percentage of people are unable to enroll in retina recognition systems because of the precision such systems require. People who are mute cannot use voice systems, and people lacking fingers or hands from congenital disease, surgery, or injury cannot use fingerprint or hand geometry systems. Although between 1 and 3 percent of the general public does not have the body part required for using any one biometric system, they are normally not counted in a system’s FTER.

Strength & Weaknesses of different BTs

Various biometric technologies have been described. There is no single biometric technique that outperforms all of them. Every technique has its own merits and demerits and usage of any technology is context dependent. Below we are giving strengths and weakness of each technology for comparison.

Strengths of Facial Recognition

• Effective for surveillance applications.


• Provides a first level “scan” within an extremely large, low-security situation.


• Easy to deploy, can use standard CCTV hardware integrated with face recognition software.


• Passive technology does not require user cooperation and works from a distance.


• May be able to use high quality images in an existing database.

Weaknesses of Facial Recognition

• Lighting, age, glasses, and head/face coverings all impact false reject rates.


• Even in surveillance applications, lower accuracy results in multiple candidates return in large populations. As a result, secondary processing is required for surveillance operations.


• Privacy concerns: people do not always know when their picture/image is being taken and being searched in a database — or worse, being enrolled in a database. Can be used without explicit opt-in permission

Strengths of Fingerprints

• Widely accepted by civil law enforcement and forensic government applications (the AFIS database); as such, fingerprints are excellent for background checks.


• Can provide a relatively low false rejection rate and false acceptance rate when used in populations with a low incidence of “outliers” (however, large groups or groups of varied by race and gender are an issue).


• Wide range of vendors and solutions.


• Ability to enroll multiple fingers.

Weaknesses of Fingerprints

• Fingerprint is not as accurate as iris recognition


• Fingerprint false accept rate varies by vendor, and is approximately 1 in 100,000.


• Iris recognition false accepts rate is 1 in 1.2 million statistically.


• Most high-end fingerprint systems measure approximately 40-60 characteristics; iris recognition looks at about 240 characteristics.

Strengths of Hand Geometry

• Currently being used for functions such as access control, employee time recording and point of sale applications.


• Fairly easy to use.


• Reasonably high acceptance among users and it is opt-in.


• Works in challenging environments.

Weaknesses of Hand Geometry

• Does not support 1: all matching with large databases.


• Weather, temperature and medical conditions such as pregnancy or certain medications can affect hand size.


• Hand size and geometry changes over time, especially in the very young and the very old.


• People are reluctant to place hand where many others have touched (hygiene issue).


• Extreme sizes are not accommodated in all hand readers.


• Fairly expensive and large equipment is required.

Strengths of Iris Recognition

• Hands-free operation.


• Proven highest accuracy: iris recognition had no false matches in over two million cross-comparisons, according to Biometric Product Testing Final Report


• Ability to handle very large populations at high speed: Iris recognition can handle very large 1: all searches within extremely large databases


• Convenient: all a person needs to do is look into a camera for a few seconds.


• A video image is taken which is non-invasive and inherently safe.


• The iris itself is stable throughout a person’s life (approximately from the age of one); the physical characteristics of the iris don't change with age.

Weaknesses of Iris Recognition

• Not a very user friendly.


• Can not recognize the person from the crowd like facial system.

Leading Biometric Technologies

A growing number of biometric technologies have been proposed over the past several years, but only in the past 5 years have the leading ones become more widely deployed. Some technologies are better suited to specific applications than others, and some are more acceptable to users. We will discuss the four leading and most deployed biometric technologies:

• Facial Recognition
• Fingerprint Recognition
• Hand Geometry
• Iris Recognition

Face Recognition
Face recognition technology is the least intrusive and fastest biometric technology. It works with the most obvious individual identifier – the human face. Instead of requiring people to place their hand on a reader or precisely position their eye in front of a scanner, face recognition systems unobtrusively take pictures of people's faces as they enter a defined area. There is no intrusion or delay, and in most cases the subjects are entirely unaware of the process. They do not feel "under surveillance" or that their privacy has been invaded.Face TechnologyFace technology is based on neural computing and combines the advantages of elastic and neural networks. Neural computing provides technical information processing methods that are similar to the way information is processed in biological systems, such as the human brain. They share some key strength, like robustness fault-resistance and the ability to learn from examples. Elastic networks can compare facial landmarks even if images are not identical, as is practically always the case in real-world situations. Neural networks can learn to recognize similarities through pattern recognition.


Face recognition is also very difficult to fool. It works by comparing facial landmarks - specific proportions and angles of defined facial features - which cannot easily be concealed by beards, eyeglasses or makeup.
Every face has numerous, distinguishable landmarks, the different peaks and valleys that make up facial features. Face recognition defines these landmarks as nodal points. Each human face has approximately 80 nodal points. Some of these measured by the software are:

• Distance between the eyes
• Width of the nose
• Depth of the eye sockets
• The shape of the cheekbones
• The length of the jaw line

These nodal points are measured creating a numerical code, called a faceprint, representing the face in the database.

Finger Print Recognition
Fingerprint recognition is one of the best known and most widely used biometric technologies. Automated systems have been commercially available since the early 1970s, and at the time of our study, we found there were more than 75 fingerprint recognition technology companies. Until recently, fingerprint recognition was used primarily in law enforcement applications. Fingerprint recognition technology extracts features from impressions made by the distinct ridges on the fingertips. The fingerprints can be either flat or rolled. A flat print captures only an impression of the central area between the fingertip and the first knuckle; a rolled print captures ridges on both sides of the finger. An image of the fingerprint is captured by a scanner, enhanced, and converted into a template. Scanner technologies can be optical, silicon, or ultrasound technologies. Ultrasound, while potentially the most accurate, has not been demonstrated in widespread use.

During enhancement, “noise” caused by such things as dirt, cuts, scars, and creases or dry, wet or worn fingerprints is reduced, and the definition of the ridges is enhanced. Approximately 80 percent of vendors base their algorithms on the extraction of minutiae points relating to breaks in the ridges of the fingertips. Other algorithms are based on extracting ridge patterns.

In the biometric process of finger scanning is a curved line in a finger image. Some ridges are continuous curves, and others terminate at specific points called ridge endings. Sometimes, two ridges come together at a point called a bifurcation. Ridge endings and bifurcations are known as minutia.

The number and locations of the minutiae vary from finger to finger in any particular person, and from person to person for any particular finger (for example, the index finger on the left hand). When a set of finger images is obtained from an individual, the number of minutiae is recorded for each finger. The precise locations of the minutiae are also recorded, in the form of numerical coordinates, for each finger. The result is a function that can be entered and stored in a computer database. A computer can rapidly compare this function with that of anyone else in the world whose finger image has been scanned.

Hand Geometry
Hand geometry systems have been in use for almost 30 years for access control to facilities ranging from nuclear power plants to day care centers. Hand geometry technology takes 96 measurements of the hand, including the width, height, and length of the fingers; distances between joints; and shapes of the knuckles. Hand geometry systems use an optical camera and light-emitting diodes with mirrors and reflectors to capture two orthogonal two-dimensional images of the back and sides of the hand. Although the basic shape of an individual’s hand remains relatively stable over his or her lifetime, natural and environmental factors can cause slight changes.

The image acquisition system comprises of a light source, a camera, a single mirror and at surface (with pegs on it). The user places his hand - palm facing downwards - on the surface of the device. The five pegs serve as control points for appropriate placement of the right hand of the user. The device also has knobs to change the intensity of the light source and the focal length of the camera. The lone mirror projects the side-view of the user's hand onto the camera.

The hand geometry-based authentication system relies on geometric invariants of a human hand. Typical features include length and width of the fingers, aspect ratio of the palm or fingers, thickness of the hand. Normally there are 14 axes along which the various features mentioned above have been measured. The five pegs on the image serve as control points and assist in choosing these axes. The hand is represented as a vector of the measurements selected above. Since the positions of the five pegs are fixed in the image, no attempt is made to remove these pegs in the acquired images.

Machine comes in compact form that can be attached with the wall quite easily.

Iris Recognition
Iris recognition technology is based on the distinctly colored ring surrounding the pupil of the eye. Made from elastic connective tissue, the iris is a very rich source of biometric data, having approximately 266 distinctive characteristics. These include the orbicular meshwork, a tissue that gives the appearance of dividing the iris radically, with striations, rings, furrows, a corona, and freckles. Iris recognition technology uses about 173 of these distinctive characteristics. Formed during the 8 months of gestation, these characteristics reportedly remain stable throughout a person’s lifetime, except in cases of injury. Iris recognition can be used in both verification and identification systems. Iris recognition systems use a small, high-quality camera to capture a black and white, high-resolution image of the iris. The systems then define the boundaries of the iris, establish a coordinate system over the iris, and define the zones for analysis within the coordinate system.

Iris patterns become interesting as an alternative approach to reliable visual recognition of persons when imaging can be done at distances of less than a meter, and especially when there is a need to search very large databases without incurring any false matches despite a huge number of possibilities. Although small (11 mm) and sometimes problematic to image, the iris has the great mathematical advantage that its pattern variability among different persons is enormous. In addition, as an internal (yet externally visible) organ of the eye, the iris is well protected from the environment and stable over time. As a planar object its image is relatively insensitive to angle of illumination, and changes in viewing angle cause only affine transformations; even the non affine pattern distortion caused by papillary dilation is readily reversible. Finally, the ease of localizing eyes in faces, and the distinctive annular shape of the iris, facilitates reliable and precise isolation of this feature and the creation of a size-invariant representation.

The Process...

Enrollment
In enrollment, a biometric system is trained to identify a specific person. The person first provides an identifier, such as an identity card. The biometric is linked to the identity specified on the identification document. He or she then presents the biometric (e.g., fingertips, hand, or iris) to an acquisition device. The distinctive features are located and one or more samples are extracted, encoded, and stored as a reference template for future comparisons. Depending on the technology, the biometric sample may be collected as an image, a recording, or a record of related dynamic measurements. How biometric systems extract features and encode and store information in the template is based on the system vendor’s proprietary algorithms. Template size varies depending on the vendor and the technology. Templates can be stored remotely in a central database or within a biometric reader device itself; their small size also allows for storage on smart cards or tokens. Minute changes in positioning, distance, pressure, environment, and other factors influence the generation of a template, making each template likely to be unique, each time an individual’s biometric data are captured and a new template is generated. Consequently, depending on the biometric system, a person may need to present biometric data several times in order to enroll. Either the reference template may then represent an amalgam of the captured data or several enrollment templates may be stored. The quality of the template or templates is critical in the overall success of the biometric application. Because biometric features can change over time, people may have to re-enroll to update their reference template. Some technologies can update the reference template during matching operations. The enrollment process also depends on the quality of the identifier presents. The reference template is linked to the identity specified on the identification document. If the identification document does not specify the individual’s true identity, the reference template will be linked to a false identity.

Verification
In verification systems, the step after enrollment is to verify that a person is who he or she claims to be (i.e., the person who enrolled). After the individual provides whatever identifier he or she enrolled with, the biometric is presented, which the biometric system captures, generating a trial template that is based on the vendor’s algorithm. The system then compares the trial biometric template with this person’s reference template, which was stored in the system during enrollment, to determine whether the individual’s trial and stored templates match. Verification is often referred to as 1:1 (one-to-one) matching. Verification systems can contain databases ranging from dozens to millions of enrolled templates but are always predicated on matching an individual’s presented biometric against his or her reference template. Nearly all verification systems can render a match–no-match decision in less than a second. A system that requires employee to authenticate their claimed identities before granting them access to secure buildings or to computers is a verification application.

Identification
In identification systems, the step after enrollment is to identify who the person is. Unlike verification systems, no identifier need be provided. To find a match, instead of locating and comparing the person’s reference template against his or her presented biometric, the trial template is compared against the stored reference templates of all individuals enrolled in the system. Identification systems are referred to as 1:N (one-to-N, or one-to-many) matching because an individual’s biometric is compared against multiple biometric templates in the system’s database. There are two types of identification systems: positive and negative. Positive identification systems are designed to ensure that an individual’s biometric is enrolled in the database. The anticipated result of a search is a match. A typical positive identification system controls access to a secure building or secure computer by checking anyone who seeks access against a database of enrolled employees. The goal is to determine whether a person seeking access can be identified as having been enrolled in the system. Negative identification systems are designed to ensure that a person’s biometric information is not present in a database. The anticipated result of a search is a non match. Comparing a person’s biometric information against a database of all who are registered in a public benefits program, for example, can ensure that this person is not “double dipping” by using fraudulent documentation to register under multiple identities. Another type of negative identification system is a surveillance system that uses a watch list. Such systems are designed to identify people on the watch list and alert authorities for appropriate action. For all other people, the system is to check that they are not on the watch list and allow them normal passage. The people whose biometrics is in the database in these systems may not have provided them voluntarily. For instance, for a surveillance system, the biometrics may be faces captured from mug shots provided by a law enforcement agency. No match is ever perfect in either verification or an identification system, because every time a biometric is captured, the template is likely to be unique. Therefore, biometric systems can be configured to make a match or no-match decision, based on a predefined number, referred to as a threshold that establishes the acceptable degree of similarity between the trial template and the enrolled reference template. After the comparison, a score representing the degree of similarity is generated, and this score is compared to the threshold to make a match or no-match decision. Depending on the setting of the threshold in identification systems, sometimes several reference templates can be considered matches to the trial template, with the better scores corresponding to better matches.

How Biometric Technology Work ?

Biometric technologies vary in complexity, capabilities, and performance, but all share several elements. Biometric identification systems are essentially pattern recognition systems. They use acquisition devices such as cameras and scanning devices to capture images, recordings, or measurements of an individual’s characteristics and computer hardware and software to extract, encode, store, and compare these characteristics. Because the process is automated, biometric decision-making is generally very fast, in most cases taking only a few seconds in real time. Depending on the application, biometric systems can be used in one of two modes: verification or identification. Verification—also called authentication—is used to verify a person’s identity—that is, to authenticate that individuals are who they say they are. Identification is used to establish a person’s identity—that is, to determine who a person is. Although biometric technologies measure different characteristics in substantially different ways, all biometric systems involve similar processes that can be divided into two distinct stages: enrollment and verification or identification.

What is Biometric ?

When used for personal identification, biometric technologies measure and analyze human physiological and behavioral characteristics. Identifying a person’s physiological characteristics is based on direct measurement of a part of the body—fingertips, hand geometry, facial geometry, and eye retinas and irises. The corresponding biometric technologies are fingerprint recognition, hand geometry, and facial, retina, and iris recognition. Identifying behavioral characteristics is based on data derived from actions, such as speech and signature, the corresponding biometrics being speaker recognition and signature recognition. Biometrics can theoretically be very effective personal identifiers because the characteristics they measure are thought to be distinct to each person. Unlike conventional identification methods that use something you have, such as an identification card to gain access to a building, or something you know, such as a password to log on to a computer system, these characteristics are integral to something you are. Because they are tightly bound to an individual, they are more reliable, cannot be forgotten, and are less easily lost, stolen, or guessed.