Biometrics Technology Centre (BTC) Biometrics Solution for Authentication Prof. David Zhang Director Biometrics Technology Centre (UGC/CRC) Department of Computing Hong Kong Polytechnic University BTC - 1 Biometrics Technology Centre (BTC) Outline Introduction to Biometrics Fingerprint Recognition Iris Recognition Speaker Recognition Signature Recognition Face Recognition Hand Recognition Applications Future Directions BTC - 2 Biometrics Technology Centre (BTC) Introduction Why Biometrics? BTC - 3 Biometrics Technology Centre (BTC) Questions In the modern world, there is an ever growing need to identify individuals. Should this person be given access to a secure system? Does this person have authorization to perform a given transaction? Is this person a citizen of our country? Every day, questions like these which are related to establishing the identity of individuals are asked millions times. BTC - 4 Biometrics Technology Centre (BTC) Current System Person have to use more and more IDs Possession (Key, Credit card, …) Knowledge (password, PIN, …) BTC - 5 Biometrics Technology Centre (BTC) Challenge None of these problems would exist if we could always be certain people - Who they say they are & the documents they use are not counterfeit. Positive identification is crucial! There are more and more highly fraudulent technologies in today’s complex society . Hence, the problem of identifying a person continues to pose a great challenge. BTC - 6 Biometrics Technology Centre (BTC) Biometrics It can overcome some of the limitations of traditional technologies because - physically present at the point-ofidentification, i.e., “something which you are or you do” Biometrics can provide better security , increased efficiency, and improved accuracy. Biometrics-based identification is emerging as the most reliable method. BTC - 7 Biometrics Technology Centre (BTC) Introduction What Biometrics? BTC - 8 Biometrics Technology Centre (BTC) Introduction Biometrics A measurable physical characteristics or personal behavioral trait used to recognize the identity, or verify the claimed identity, of an enrollee BTC - 9 Biometrics Technology Centre (BTC) Examples of Biometrics BTC - 10 Biometrics Technology Centre (BTC) Characteristics of Biometrics Universality Uniqueness Collectability Permanence Acceptability Circumvention BTC - 11 Biometrics Technology Centre (BTC) Taxonomy of Biometrics Biometrics Physical Head Hand Behavioral Signature Voice Others (Gesture, gait, Keystroke) Face Iris Hand Geometry Others (Ear, Retina, Dental) Finger print Palm print Others (finger) Taxonomy by Technology Type BTC - 12 Biometrics Technology Centre (BTC) Introduction How Biometrics? BTC - 13 Biometrics Technology Centre (BTC) Biometrics Based Identify Procedure Sensor Measures of biometrics Image & Signal Processing Pattern Recognition Extraction of features Comparing the features Decision Theory Making the decision BTC - 14 Biometrics Technology Centre (BTC) Biometrics Evaluation BTC - 15 Biometrics Technology Centre (BTC) Demo Demo DEMO DEMO BTC - 16 Biometrics Technology Centre (BTC) BTC - 17 Biometrics Technology Centre (BTC) Fingerprint Development Current State Stability and uniqueness: Based on a century of examination, it is estimated that the chance of two people, including twin, having the same print is less than one on a billion. Fingerprint identification is the most widespread application of biometrics technology. First commercial ABS was used in 1971. Feature Set Minutiae Endpoint & junction of print ridges, and position, direction & relation between them Singular Point BTC - 18 Biometrics Technology Centre (BTC) Demo Demo BTC - 19 Biometrics Technology Centre (BTC) BTC - 20 Biometrics Technology Centre (BTC) Iris Development Current State Iris features can be more easily obtained than other features from eyes, like retina. It is shown that an iris has more details than a fingerprint. Highly detailed and unique texture will remain stable over decades of life. Feature Set Textures with striations, contraction furrows, pits, collagenous fibers, filament, crypts (darkened areas resembling excavations), serpentine vasculature, rings, and freckles BTC - 21 Biometrics Technology Centre (BTC) Demo De mo BTC - 22 Biometrics Technology Centre (BTC) BTC - 23 Biometrics Technology Centre (BTC) Palmprint Development Current State It has relatively stable physical characteristics, basically unalterable without trauma to the individual. Many useful features extracted from both fingerprint & hand shape can be simultaneously obtained. However, no palmprint system is currently used in automatic biometric identification. Therefore, it is necessary to develop a new approach to automated personal identification by palmprint. Feature Set Minutiae (like fingerprint) and shape (like hand) Principal line, wrinkle, delta point and region feature BTC - 24 Biometrics Technology Centre (BTC) Demo BTC - 25 Biometrics Technology Centre (BTC) BTC - 26 Biometrics Technology Centre (BTC) Speech Development Current State The least invasive of the biometric recognition technologies and the most natural to use is speech system. Just say a phrase, about a second long - any language or dialect - chosen by the user. A typical case is AT&T Smart Card. Feature Set Twenty parameters – Pits, speed, energy, density, and waveform BTC - 27 Biometrics Technology Centre (BTC) Demo Demo BTC - 28 Biometrics Technology Centre (BTC) BTC - 29 Biometrics Technology Centre (BTC) Signature Development Current State It should distinguish between person’s habitual parts and those that vary with almost every signing. Two methods: on-line & off-line, where wired pens & sensitive tables are need for on-line signature. Feature Set Off-line – Line segments H line, V line, P line, N line, C curve, D curve, A curve, V curve, S curve, Z curve, curve and circle On-line Pen tip position (x, y), distance above the surface (h), pressure (p), tilt angle (tx, ty) and velocity (vx, vy) BTC - 30 Biometrics Technology Centre (BTC) Demo BTC - 31 Biometrics Technology Centre (BTC) BTC - 32 Biometrics Technology Centre (BTC) Face Development Current State Using the whole face for automatic identification is a complex task because its appearance is constantly changing. One effective approach may employ rule-based logic and a neural network for the image classification process. The first face system is introduced in 1992. Feature Set Facial geometry Size of eye, distance from eye to mouth, middle of mouth to chin, side of eye to cheek, size of mouth, radius vectors and feature points Face template BTC - 33 Biometrics Technology Centre (BTC) Demo DEMO BTC - 34 Biometrics Technology Centre (BTC) BTC - 35 Biometrics Technology Centre (BTC) Hand Development Current State It is a simple ABS, which features are obtained by three-dimensional shape. Analysis is based on measurement and comparison of geometries. As a biometric application, one hand geometry identification device is commercially available in 1994. Feature Set Finger length, width, area, thick and their relationship between fingers BTC - 36 Biometrics Technology Centre (BTC) Demo DEMO BTC - 37 Biometrics Technology Centre (BTC) BTC - 38 Biometrics Technology Centre (BTC) Biometrics Applications Biometrics Applications Summary of Application Dental Recognition Forensics DNA Recognition Forensics, medicine, genetics Fingerprint Recognition/ Verification Face Recognition/ Verification Hand Verification Signature Recognition/ Verification Verification Immigration & naturalization, welfare distribution, military identification, forensics, access control Suspect description & identification, missing persons, licenses, credit card, welfare distribution Access control, immigration and naturalization, services distribution Signature verification, identification from handwriting Access control Recognition/ Verification Speaker verification, phone service, speaker verification, access control Iris Voice BTC - 39 Biometrics Technology Centre (BTC) Biometrics in Banking Application Banking Services Biometrics Bankbook Face Credit Card Hand/Finger IRIS Telephone Banking Fingerprint Intelnet Banking Palm print Cheque Voice Safe Deposit Box Signature BTC - 40 Biometrics Technology Centre (BTC) Typical Applications Verification systems computer login cheque cashing airport check-in Identification systems Driving licences Walfare Police mug-shots BTC - 41 Biometrics Technology Centre (BTC) EXAMPLE 1: Access Control DEMO BTC - 42 Biometrics Technology Centre (BTC) EXAMPLE 2: Law Enforcement BTC - 43 Biometrics Technology Centre (BTC) EXAMPLE 3: Smart Card DEMO BTC - 44