Introduction to Biometrics

1. Introduction to Biometrics Dr. Bhavani Thuraisingham The University of Texas at Dallas Lecture #7 Biometric Technologies: Finger Scan September 14, 2005

2. Outline • Introduction • Basic Terms • Technologies • Finger Scan Process • Feature Extraction • Classification • Accuracy and Integrity • Biometric Middleware • Strengths and Weaknesses • Biometric vs. Non Biometric Fingerprinting • Research Directions • Project Related Information

3. References • Course Text Book, Chapter 4 • http://www.biometricsinfo.org/fingerprintrecognition.htm

4. Introduction • What is Finger-Print Scanning • Fingerprint scanning is the acquisition and recognition of a person’s fingerprint characteristics for identification purposes. • This allows the recognition of a person through quantifiable physiological characteristics that verify the identity of an individual. • Methods • There are basically two different types of finger-scanning technology that make this possible. • One is an optical method, which starts with a visual image of a finger. • The other uses a semiconductor-generated electric field to image a finger.

5. Introduction (Concluded) • There are a range of ways to identify fingerprints. • traditional police methods of matching minutiae • straight pattern matching • Ultrasonics • Fingerprint revenues are projected to grow from $144.2m in 2002 to $1,229.8m in 2007. Fingerprint revenues are expected to comprise approximately 30% of the entire biometric technologies • Applications • to access networks and PCs, enter restricted areas, and to authorize transactions. • Deployed in many locations (discussed in text book)

6. Basic Terms • Components • Image acquisition systems, image processing components, template generation and matching components, storage components • Surface on which finger is placed is Platen or Scanner • Finger scan module • consists of platen + printed circuit board + standard connector that transmits digitized information to a peripheral or standalone device

7. Example Technologies • Optical Technology • Oldest technology • Camera registers the image of the fingerprint against a coated glass or plastic platen • Black, gray and white lines • Silicon Technology • Silicon chip embedded in a platen • High image quality • Commercially available since around 1998 • Ultrasound Technology • Transmit acoustic waves to the finger and generating images

8. Process • Image Acquisition • Measured in terms of dots per inch • Center of the finger print must be placed on the platen • Need appropriate size for platen • Image Processing • Eliminate gray areas from image • Convent gray pixels to black and white pixels • Location of Distinctive Characteristics • Fingerprints consists of ridges and valleys • Swirls, loops, arches, deltas • Ridges and valleys are characterized by irregularities called minutiae • A finger scan image can produce about 15-50 minutiae

9. Process (Concluded) • Template Creation • Vendors use proprietary algorithms • Depends on the following • Location and angle of a minutiae point • Distance and position of minutiae relative to the core • Type and quality of the minutiae • Need to eliminate sweat, scars, dirt, etc. • Template matching • May depend on the number of minutiae matched

10. Methods of Finger PrintingMinutiae vs. Pattern matching • Minutiae • Most of the finger-scan technologies are based on minutiae • Pattern Matching • Feature extraction and template generation based on series of ridges as opposed to discrete points • Advantage: Minutiae points affected by wear and tear • Disadvantage: Sensitive to proper placement of finger; large storage for templates • Correlation • Michigan State University of developing correlation based methods

11. Feature Extraction • The human fingerprint is comprised of various types of ridge patterns • left loop, right loop, arch, whirl, and tented arch. • Loops make up nearly 2/3 of all fingerprints • whirls are nearly 1/3 • 5-10% are arches. • Figure 1 • Source: Book, URL

12. Feature Extraction (Continued) • Minutiae (Figure 1), the discontinuities that interrupt the otherwise smooth flow of ridges, are the basis for most fingerprint authentication. • Many types of minutiae exist, including dots (very small ridges), islands (ridges slightly longer than dots), ponds or lakes - - - - • The core is the inner point, normally in the middle of the print, around which swirls, loops, or arches center. • Deltas are the points, normally at the lower left and right hand of the fingerprint, around which a triangular series of ridges center. • The ridges are also marked by pores, which appear at steady intervals.

13. Feature Extraction (Continued) • Once a high-quality image is captured, there are a several steps required to convert its distinctive features into a compact template. • This process, known as feature extraction, is at the core of fingerprint technology. • fingerprint vendor has a proprietary feature extraction mechanism • The image must then be converted to a usable format. • If the image is grayscale, areas lighter than a particular threshold are discarded, and those darker are made black • The ridges are then thinned from 5-8 pixels in width down to one pixel, for precise location of endings and bifurcations.

14. Feature Extraction (Continued) • Minutiae localization begins with this processed image. • At this point, even a very precise image will have distortions and false minutiae that need to be filtered out • an algorithm may search the image and eliminate one of two adjacent minutiae, as minutiae are very rarely adjacent. • Anomalies caused by scars, sweat, or dirt appear as false minutiae, and algorithms locate any points or patterns that do not make sense • A large percentage of would-be minutiae are discarded in this process.

15. Feature Extraction (Concluded) • The point at which a ridge ends, and the point where a bifurcation begins, are the most rudimentary minutiae, and are used in most applications. • There is variance in how exactly to situate a minutia point: • whether to place it directly on the end of the ridge, one pixel away from the ending, or one pixel within the ridge ending • Once the point has been situated, its location is commonly indicated by the distance from the core, with the core serving as the 0,0 on an X,Y-axis. • Some vendors classify minutia by type and quality. The advantage of this is that searches can be quicker

16. Fingerprint Classification • Large volumes of fingerprints are collected and stored everyday in a wide range of applications including forensics, access control, and driver license registration. • An automatic recognition of people based on fingerprints requires that the input fingerprint be matched with a large number of fingerprints in a database (FBI database contains approximately 70 million fingerprints). • To reduce the search time and computational complexity, it is desirable to classify these fingerprints in an accurate and consistent manner so that the input fingerprint is required to be matched only with a subset of the fingerprints in the database.

17. Fingerprint Classification (Continued) • Fingerprint classification is a technique to assign a fingerprint into one of the several pre-specified types already established in the literature which can provide an indexing mechanism. • Fingerprint classification can be viewed as a coarse level matching of the fingerprints. • An input fingerprint is first matched at a coarse level to one of the pre-specified types and then, at a finer level, it is compared to the subset of the database containing that type of fingerprints only.

18. Fingerprint Classification (Concluded) • Michigan State University has developed an algorithm to classify fingerprints into five classes, • whirl, right loop, left loop, arch, and tented arch. • The algorithm separates the number of ridges present in four directions (0 degree, 45 degree, 90 degree, and 135 degree) by filtering the central part of a fingerprint with a bank of Gabor filters. • This information is quantized to generate a FingerCode which is used for classification. • Classification is based on a two-stage classifier which uses a K-nearest neighbor classifier in the first stage and a set of neural networks in the second stage. • The classifier is tested on 4,000 images in the NIST-4 database with about 90% accuracy

19. Image Enhancement • A critical step in automatic fingerprint matching is to automatically and reliably extract minutiae from the input fingerprint images. • However, the performance of a minutiae extraction algorithm relies heavily on the quality of the input fingerprint images. • In order to ensure that the performance of an automatic fingerprint identification/verification system will be robust with respect to the quality of the fingerprint images, it is essential to incorporate a fingerprint enhancement algorithm in the minutiae extraction module. • Michigan State University has developed algorithms for enhancement

20. Image Enhancement (Concluded) Source: URL

21. Accuracy and Integrity • In most cases, false negatives (a failure to recognize a legitimate user) are more likely than false positives. • Overcoming a fingerprint system by presenting it with a "false or fake" fingerprint will be difficult • Sensors on the market use a variety of means to circumvent the problems. • Problem: Someone may attempt to use latent print residue on the sensor just after a legitimate user accesses the system; Presenting a finger to the system that is no longer connected to its owner. • Solutions: Sensors attempt to determine whether a finger is live, and not made of latex; Detectors for temperature, blood-oxygen level, pulse, blood flow, humidity, or skin conductivity would be integrated.

22. Biometric Middleware • Enables various biometric technologies • Allows match / no-match decisions made by core technologies to provide authentication to various applications • Similar to the concept of middleware in systems • Integrates applications and resources • Flexible middleware • Solutions adapted for applications

23. Biometric Middleware (Concluded) Immigration Entry Building Entry PC Entry Middleware Applications Middleware Middleware Services Iris-scan Face-scan Finger-scan

24. Strengths and Weaknesses • Strengths • Proven technology and high level of accuracy • Many deployments • Easy to use • Can enroll multiple fingers • Weaknesses • Some users do not have clear fingerprints • Over time image quality deteriorates • Privacy concerns

25. Biometric vs. Non Biometric Fingerprinting • Fingerprinting, is the acquisition and storage of the image of the fingerprint. • Two types of systems • Forensic (AFIS – Automatic Fingerprint Identification System) • Biometric system • AFIS stores images of fingerprints; need large amount if storage. • Biometric systems store particular data about the fingerprint in a much smaller template. • After the data is extracted, the fingerprint is not stored • The full fingerprint cannot be reconstructed from the fingerprint template. • Used to log on to PC

26. Biometric vs. Non Biometric Fingerprinting (Concluded) • Response time - AFIS systems may take hours to match a candidate, while fingerprint systems respond within seconds • Cost - an AFIS capture device is very expensive.  A PC peripheral fingerprint device is much cheaper • Accuracy - an AFIS system might return the top 5 candidates with the intent of locating or questioning the top suspects. Fingerprint systems are designed to return a single yes/no answer. • Scale – AFIS systems scalable to thousands and millions of users. Fingerprint systems are and do not require significant processing power. • Capture – AFIS systems are designed to use the entire fingerprint. Fingerprint systems use only the center of the fingerprint. • Storage – AFIS systems need large storage. Fingerprint systems do not • Infrastructure – AFIS systems require a backend infrastructure for storage, matching, and duplicate resolution. Fingerprint systems rely on a PC or a peripheral device for processing and storage.

27. Some Research Directions • New Biometric Technologies • Less False Positives and False Negatives • Better Performance • Secure Biometrics • Privacy • Societal Impact

28. Summary • Most popular biometric technology • Fairly high accuracy • Market expected to grow a great deal • Feature extraction is the key mechanism • Minutiae based and non Minutiae based methods for Biometric matching • Differences between systems used for forensic applications and biometric systems

29. Some Project Related Information Dr. Bhavani Thuraisingham The University of Texas at Dallas Graduate Student: Pallabi Parveen September 14, 2005

30. TA Office Hours • Nathalie Tsybulnik • 7-10pm Monday in ECSS 3.403 • Tuesday from 10.00-11.00am she will usually be in the general lab downstairs.

31. Some Tools for Project • http://java.sun.com/products/java-media/jai/forDevelopers/jaifaq.html#what • Java Advanced Imaging Toolkit (product of Sun Microsystems) • Can Download • http://www.mathworks.com/products/image/ • Matlab Image Processing • Matlab available in some of our labs • Cannot download • CMU Voice Recognition Open Source System Sphinx • http://cmusphinx.sourceforge.net/html/cmusphinx.php

32. Face Recognition • Given at CMU, involves face recognition using neural networks. • 32 images of each of 20 students in the class were taken with a variety of head positions and facial expressions. • These images were then used to train and test neural networks to recognize individual people, and to recognize different face • Source: • http://www.cs.cmu.edu/afs/cs.cmu.edu/user/avrim/www/ML94/face_homework.html

33. Finger Print Recognition • Fingerprint Minutiae from Latent and Matching Tenprint Images • NIST Special Database 27 contains latent fingerprints from crime scenes and their matching rolled fingerprint mates. • Source: http://www.nist.gov/srd/nistsd27.htm • http://www.itl.nist.gov/iad/894.03/databases/defs/dbases.html#finglist\

34. Fingerprint Research • NIST 8-Bit Gray Scale Images of Fingerprint Image Groups (FIGS) • 2000 8-bit gray scale fingerprint image pairs including classifications • 400 fingerprint pairs from each of the five classifications - Arch, Left and Right Loops, Tented Arch, Whirl) • Source: http://www.nist.gov/srd/nistsd4.htm

35. Iris Recognition • CASIA Iris Image Database( ver 1.0) includes 756 iris images from 108 eyes (hence 108 classes). • For each eye, 7 images are captured in two sessions, where three samples are collected in the first session and four in the second session. • Source: http://www.sinobiometrics.com/casia%20iris.htm

36. Keystroke Dynamics as a Biometric for Authentication • An emerging non-static biometric technique that aims to identify users based on analyzing habitual rhythm patterns in the way they type [Fabian Monrose et al.]. • Source: http://www.cs.jhu.edu/~fabian/papers/fgcs.pdf