Introduction to Biometrics Dr Bhavani Thuraisingham The University

Introduction to Biometrics Dr. Bhavani Thuraisingham The University of Texas at Dallas Lecture #13 Biometric Technologies: Some Physiological Biometrics October 5, 2005

Outline l Summary of Previous Lectures on Biometrics Technologies l Some other biometrics

References l Course Text Book, Chapter l http: //www. biometricsinfo. org/ l http: //ctl. ncsc. dni. us/biomet%20 web/BMRetinal. html l http: //www. howstuffworks. com/dna-evidence. htm

Summary of Previous Lectures on Biometrics Technologies l Fingerprint Scan l Face Scan l Iris Scan l Voice Scan

Some Other Biometrics Technologies l Hand Scam l Retina Scan l AFIS (Automated Finger print Identification System) l Multimodal Biometrics l DNA Biometrics l Some Behavioral Biometrics - Signature recognition - Keystroke Dynamics

Hand Scan: Introduction l This biometric approach uses the geometric form of the hand for confirming an individual’s identity. l Because human hands are not unique, specific features must be combined to assure dynamic verification. l Some hand-scan devices measure just two fingers, others measure the entire hand. l Features include characteristics such as finger curves, thickness and length; the height and width of the back of the hand; the distances between joints and all bone structure. l Although the bone structure and joints of a hand are relatively constant traits, other influences such as swelling or injury can disguise the basic structure of the hand.

Hand Scan: Introduction (Concluded) l To register in a hand-scan system a hand is placed on a reader’s covered flat surface. l This placement is positioned by five guides or pins that correctly situate the hand for the cameras. l A succession of cameras captures 3 -D pictures of the sides and back of the hand. l The hand-scan device can process the 3 -D images in 5 seconds or less and the hand verification usually takes less than 1 second. l Components include: Acquisition hardware, Matching software, Storage

Hand Scan: How it works l Hand geometry scanners such as those made by Recognition Systems Inc. take over 90 measurements of the length, width, thickness, and surface area of the hand four fingers--all in just 1 second. l The technology uses a 32, 000 -pixel CCD digital camera to record the hand's three-dimensional shape from silhouetted images projected within the scanner. l The scanner disregards surface details, such as fingerprints, lines, scars, and dirt, as well as fingernails, which may grow or be cut from day to day. l When a person uses the scanner, it compares the shape of the user's hand to a template recorded during an enrollment session. If the template and the hand match, the scanner produces an output--it may unlock a door, transmit data to a computer, verify identification, or log the person's arrival or departure time.

Hand Scan: How it works l During enrollment, which takes approximately 30 seconds, the user places the right hand in the reader three times. The unit's internal processor and software convert the hand image to a 9 -byte mathematical template, which is the average of the three readings. l The user's template may reside in internal memory (capable of holding over 27, 000 users), or on other media such as a hard disk or smart card chip. l As opposed to such technologies as fingerprint, voice recognition, and facial recognition, where a multitude of vendors compete via their proprietary technology, hand geometry technology is dominated by one company, Recognition Systems, Inc. (RSI) l Finger geometry is led by Biomet Partners.

Hand Scan: How it works (Continued) l RSI's method for capturing the biometric sample is as follows: To enroll, the users places his or her hand palm down on the reader's surface. l The user then aligns his or her hand with the five pegs designed to indicate the proper location of the thumb, forefinger, and middle finger. l Three placements are required to enroll on the unit; the enrollment template is a representation of the most relevant data from the three placements. l RSI's units use a 32, 000 -pixel CCD (charged coupled device) digital camera, inferring the length, width, thickness, and surface area of the hand fingers from silhouetted images projected within the scanner.

Hand Scan: How it works (Concluded) l Over 90 measurements are taken, and the hand fingers' characteristics are represented as a 9 byte template. source: Recognition Systems, Inc. l Biomet Partners' technology is similar, but draws on the shape and characteristics of the index and middle finger. The data is saved as a 20 byte template. l Hand geometry does not perform 1 -to-many identification, as similarities between hands are not uncommon. l Where hand geometry does have an advantage is in its FTE (failure to enroll) rates, which measure the likelihood that a user is incapable of enrolling in the system. Fingerprint, by comparison, is prone to FTE's due to poor quality fingerprints; facial recognition requires consistent lighting to properly enroll a user.

Hand Scan: Template Generation and Matching l Distinctive features include height, width, thickness of the hand l Distinctive features of the hand finger are extracted from a series of 3 -D images and recorded into a small templates l False matching and false non-matching are possible due to the fact that hands may swell and undergo changes

Hand Scan: Applications l Hand geometry is currently among the most widely used biometric technologies, most suitable for access control and time and attendance applications. l Hand scan is used reliably at thousands of places of employment, universities, apartment buildings, and airports - anyplace requiring reasonably accurate, non-intrusive authentication. l The nature of hand geometry technology is such that most projects are fairly small-scale and involve only a handful of readers, but there are some projects which incorporate dozens of readers.

Hand Scan: Deployments l INSPASS (Immigration and Naturalization Service Passenger Accelerated Service System) project, one which allows frequent travelers to circumvent long immigration lines at international airports. l Qualified passengers, after enrolling in the service, receive a magstripe card encoded with their hand scan information. Instead of being processed by passport control personnel, INSPASS travelers swipe their card, place their hand, and proceed with their I-94 to the customs gate. l Nearly 50, 000 people have enrolled in the service, and approximately 20, 000 verifications take place every month. Travelers from 30 different countries are qualified to register for INSPASS; pending budgetary constraints, the near-term objective is to rollout the INSPASS project to over 20 airports in the U. S.

Hand Scan: Market Size l Hand geometry is projected to be one of the slowed growing biometric technology through 2007. l Because the range of applications in which hand geometry is typically limited to access control and time and attendance, it will draw a progressively smaller percentage of biometric revenues. l Overall, hand geometry revenues are projected to grow from $27. 7 m in 2002 to $97. 4 m in 2007. Hand geometry revenues are expected to comprise approximately 2. 5% of the entire biometric market.

Hand Scan: Strengths and Weakness l Strengths - Ease of use. - Resistant to fraud. - Template size - Using RSI, a template size of 9 bytes is extremely small User perceptions – non-intrusive l Weaknesses Static design - largely unchanged for years. Cost Injuries to hands Accuracy, hand geometry, in its current incarnation, cannot perform 1 -to-many searches, but instead is limited to 1 -to-1 verification. -

Retina Scan: Overview l Completely different from Iris Scan l Camera captures the image of the retina l Movements affects the images l Need about 3 – 5 images for enrollment l Distinctive features include network of blood vessels l Glaucoma and other conditions may affect retina scan l Template generation process will map the unique network of blood vessels into a template l Template is about 96 bytes l Usually does one-many identification l Good for highly secure environments

Retina Scan: Overview (Concluded) l Strengths - Resistance to false matching - Stable characteristics l Weakness - Difficult to use - User discomfort - Limited applications

Retina Scan: Details l Retinal scanning analyses the layer of blood vessels at the back of the eye. l Scanning involves using a low-intensity light source and an optical coupler and can read the patterns at a great level of accuracy. l The user looks through a small opening in the device at a small green light. The user must keep their head still and eye focused on the light for several seconds during which time the device will verify his identity. This process takes about 10 to 15 seconds total. l There is no known way to replicate a retina, and a retina from a dead person would deteriorate too fast to be useful, so no extra precautions have been taken with retinal scans to be sure the user is a living human being.

Retina Scan: Details (Continued) l Retina scan is actually one of the oldest biometrics as 1930's research suggested that the patterns of blood vessels on the back of the human eye were unique to each individual. l While technology has taken more time than theory to be usable, Eye. Dentify, founded in 1976, developed The Eyedentification 7. 5 personal identification unit, the first retina scan device made for commercial use, in 1984. l At this time, they are still the primary company for retinal scan devices l Retina scan is used almost exclusively in high-end security applications l It is used for controlling access to areas or rooms in military installations, power plants, and the like that are considered high risk security areas.

Retina Scan: Details (Concluded) l Retina scan devices are provide accurate biometric l The continuity of the retinal pattern throughout life and the difficulty in fooling such a device also make it a great longterm, high-security option. l The the cost of the proprietary hardware as well as the inability to evolve easily with new technology make retinal scan devices a bad fit for most situations. l It also has the stigma of consumer's thinking it is potentially harmful to the eye, and in general not easy to use.

AFIS l Automated Fingerprint Identification System (AFIS) technology is used in a variety of law enforcement and civil applications. l In law enforcement, fingerprints are collected from arrested subjects and searched against local, state, regional, and/or national fingerprint databases. l The subject's ten fingerprints are acquired either through the traditional ink-and-roll method or through an optical livescan system, consisting of a sizeable fingerprint scanner, PC, and imaging and transmission software l The electronic fingerprints are submitted, along with demographic data, to identify or verify the identity of the subject.

AFIS (Concluded) l Searches may take minutes, hours, or days, depending on the quality of the information submitted, the size of the database being searched, and the entity requesting the search. l Law enforcement searches often return candidate lists used to determine which of several possible matches is the best match. l Most widely used biometric technology l AFIS is different from fingerprinting systems AFIS captures (in addition to templates) and uses image analysis algorithm -

Multimodal Biometrics l A multimodal biometric system uses multiple applications to capture different types of biometrics. l This allows the integration of two or more types of biometric recognition and verification systems in order to meet stringent performance requirements. l A multimodal system could be, for instance, a combination of fingerprint verification, face recognition, voice verification and smart-card or any other combination of biometrics. l This enhanced structure takes advantage of the proficiency of each individual biometric and can be used to overcome some of the limitations of a single biometric.

DNA Biometrics l Proving that a suspect's DNA matches a sample left at the scene of a crime requires two things: Creating a DNA profile using basic molecular biology protocols; Crunching numbers and applying the principles of population genetics to prove a match mathematically l Humans have 23 pairs of chromosomes containing the DNA blueprint that encodes all the materials needed to make up your body as well as the instructions for how to run it. One member of each chromosomal pair comes from your mother, and the other is contributed by your father. l Every cell in your body contains a copy of this DNA; While the majority of DNA doesn't differ from human to human, some 3 million base pairs of DNA (about 0. 10 percent of your entire genome) vary from person to person. l The key to DNA evidence lies in comparing the DNA left at the scene of a crime with a suspect's DNA in these chromosomal regions that do differ.