Fingerprint Sensing Techniques Devices and Applications Rahul Singh

Fingerprint Sensing Techniques, Devices and Applications Rahul Singh kingtiny@cs. cmu. edu 30 th April 2003 1

Fingerprint Biometric ► First used in China in 700 AD ► Proposed in Europe in 1858, implemented in Germany in 1903. ► Unique – So far no two prints from different fingers have been found that are identical 2

Fingerprint Biometric Characteristics ► Fingerprint is the representation of the epidermis of a finger ► Set of (almost/often) parallel ridge lines ► Ridges produce local patterns Source: http: //www. biometrika. it/eng/wp_fingintro. html 3

Fingerprint Biometric Characteristics ► Five main classes of fingerprints § § § Arch Tented Arch Left Loop Right Loop Whorl Source: http: //www. biometrika. it/eng/wp_fingintro. html 4

Fingerprint Sensing ► Two stages 1. 2. 3. Capture Fingerprint image Process image and extract features Store data for comparison or compare with stored templates 5

Types of Fingerprint Sensors ► Optic Reflexive § Finger lies on a prism. Total internal reflection produces image of fingerprint on a camera chip ► Optic Transmissive with Fiber Optic Plate § Light source illuminates through the finger § Finger lies on fiber-optic plate that transmits image data to camera chip ► Optical Line § Pixel array measures the light reflected by the finger Source: http: //home. t-online. de/home/manfred. bromba/fpfaqe. htm 6

Types of Fingerprint Sensors ► Capacitive Line § Capacitor array measures the capacitance at each pixel ► Thermal Line § Finger is moved across a narrow array of thermal sensors § Temperature varies across the grooves and ridges § Thermal sensors measure the temperature differences over time ► Pressure Sensitive § Sensor measures the pressure per pixel ► Dynamic Capacitive § Capacitance is measured by A/C voltage Source: http: //home. t-online. de/home/manfred. bromba/fpfaqe. htm 7

Types of Fingerprint Sensors ► Static Capacitive Type 1 § One electrode per pixel § Capacitance measured w. r. t neighboring pixel. § If pixel is on a groove capacitance is small § If pixel is on a ridge then capacitance is large ► Static Capacitive Type 2 § Same as above except capacitance is measure w. r. t ground ► Acoustic (Ultrasound) § Image of fingerprint is recorded by very high frequency sound Source: http: //home. t-online. de/home/manfred. bromba/fpfaqe. htm 8

Capacitive Sensing ► Fingerprint consists of tightly spaced ridges and valleys ► Sensor consists of a capacitive array ► Capacitive array acts as one plate of a capacitor while the finger acts as the other ► Each pixel in the array is charged to a reference voltage and allowed to discharge with a reference current ► The rate of change of potential at each pixel is proportional to the capacitance seen by the array 9

Capacitive Sensing 1. 2. 3. 4. 5. Charge amp reset. Inverter O/P settles to threshold Ref. charge applied to I/P O/P Voltage proportional to feedback capacitance Inverter O/P = upper saturation level if there is no feedback capacitance Inverter O/P = close to logical threshold when feedback capacitance is large Source: http: //www-micro. deis. unibo. it/~tartagni/Finger. Sensor. html 10

Capacitive Sensing ► 300 x 300 pixel array (90, 000 pixels) ► 500 dpi Fingerprint image Source: http: //www. fme. fujitsu. com/products/biometric/pdf/Find_FPS. pdf 11

Optical Sensing Finger touches light emitting Tactile. Sense polymer ► Photodiode array embedded in the glass detects illumination ► Image is captured and transferred for storage ► Source: [Tactilesense] http: //www. ethentica. com/tactwhtpr. pdf 12

Optical Sensing ► Sensing by projecting an image of the fingerprint onto a camera by total internal reflection. Source: http: //www. biometrika. it/eng/wp_fingintro. html 13

Optical Vs Capacitive ► Capacitive § Greater miniaturization § Newer technology § Can be embedded into small devices § Prone to dirt etc since finger touches silicon § Relatively cheap ► Optical Sensors § Larger sensing area since manufacturing large pure silicon chips is expensive § More robust. Longer life § More expensive § Better image quality and higher resolution 14

Factors affecting the scan ► Image § § § quality Sharpness Contrast Distortion Source: http: //www. biometrika. it/eng/wp_scfing. html 15

Factors affecting the scan ► Resolution – higher is better § Too low and we cannot detect the minutiae ► Sensing area § Average fingerprint is about 0. 5” x 0. 7” § Large area (1. 0” x 1. 0”) ensures that overlap effects (leading to false rejections) are reduced Source: http: //www. biometrika. it/eng/wp_scfing. html 16

Data Storage and Matching ► Minutiae or Galton Characteristics § Termination of Ridge lines § Bifurcation of Ridge lines Source: http: //www. biometrika. it/eng/wp_fingintro. html 17

Data Storage and Matching ► Final data size = 300 to 600 bytes Source: http: //www. fme. fujitsu. com/products/biometric/pdf/Find_FPS. pdf 18

Data Storage and Matching ► Directional Map § Discrete matrix whose elements denote the orientation of the tangent to ridge lines 19

FX 2000 ► FX 2000 – Optical Sensor Efficiency § Database of 100 users (non-experts) § Low quality fingerprints TASK SPEED_DEFAULT SPEED_FAST Feature extraction 350 ms 260 ms Matching 120 ms 47 ms Identity verification (1: 1) 470 ms 307 ms Identification (1: 50) 3. 35 sec 1. 43 sec Accuracy threshold t SPEED_DEFAULT Verification time (1: 1) SPEED_FAST FAR FRR 0. 3500 0. 0049 (0. 49%) 0. 0005 (0. 05%) 0. 0032 (0. 32%) 0. 0009 (0. 09%) 0. 3750 0. 0025 (0. 25%) 0. 0010 (0. 10%) 0. 0014 (0. 14%) 0. 4000 0. 0011 (0. 11%) 0. 0014 (0. 14%) 0. 0005 (0. 05%) 0. 0020 (0. 20%) 0. 4250 0. 0006 (0. 06%) 0. 0019 (0. 19%) 0. 0002 (0. 02%) 0. 0034 (0. 34%) 0. 4500 0. 0004 (0. 04%) 0. 0026 (0. 26%) 0. 0000 (0. 00%) 0. 0049 (0. 49%) 0. 4750 0. 0000 (0. 00%) 0. 0036 (0. 36%) 0. 0000 (0. 00%) 0. 0063 (0. 63%) Time to verify the identity Identification time (1: 50) Average time to identify an individual. 50 Users. Match is found in the middle. 20

Secugen FDA 01/FCA 01 ► Optical sensor ► Resolution = 500 dpi ► Verification time = < 1 second ► Sensing area = 13. 6 mm x 16. 2 mm 21

Authentec Finger. Loc ► AF-S 2 § Capacitive § 68 pin PLCC § Resolution: 250 dpi § Array size: . 512”x. 512” ► AFS 8500 § Capacitive § 144 pin LQFP § Resolution: 250 dpi § Array Size: . 384” x. 384” 22

Biomouse/Biomouse plus ► Optical sensor ► “High speed” matching algorithm – 400 prints per second on p. II 400. ► Resolution = 500 dpi ► Average template size = 350 bytes ► Biomouse Plus comes with built in smart card reader 23

Defeating Fingerprint Scanners ► Gummi bears defeat fingerprint sensors § Japanese cryptographer § Gelatin + plastic mould § Latent fingerprints from glass § Cyanoacrylate Adhesive (superglue fules) § Digital camera § Adobe Photoshop § Photosensitive PCB – etched print in copper § Moulded finger with print Source: http: //www. theregister. co. uk/content/55/25300. html 24

Defeating Fingerprint Sensors ► More sophisticated devices use incorporate biosensing modules prior to fingerprint capture ► Detect blood flow ► Detect body heat ► Sensor shuts down if no life is detected 25

Types of attack ► Brute force ► Latent print ► Replay ► Trojan Horses ► Fake feature ► Dead feature ► Other (software leaks, bad security policies etc) 26

Applications ► Secure logins via keyboard modules ► User identification at kiosks ► Biometric door locks ► Credit card security ► Weapon activation ► Theft protection 27

Fingerprint Verification for Smart Cards Motorola, Australia ► Senior Honors thesis ► Develop biometric security solution (prototype) for Motorola dual-slot phones ► Users insert credit card into slot 1 for ecommerce ► Smart card with embedded biometric into slot 2 ► Fingerprint sensor on phone identifies user and authorizes use of credit card 28

Fingerprint Verification for Smart Cards Motorola, Australia Enrollment X. 509 Certificate Fingerprint template 1010011010100101 Smart Card X. 509 Certificate Fingerprint template 1010011010100101 http: //www. roma. unisa. edu. au/08216/99 u/index. html 29

Fingerprint Verification for Smart Cards Motorola, Australia Verification Fingerprint template 1010011010100101 Compare Smart Card X. 509 Certificate Fingerprint template 1010011010100101 http: //www. roma. unisa. edu. au/08216/99 u/index. html 30

Questions ? 31