Why Biometrics?

Biometrics is a security solution based on something you know, have, and are:

Know	Password, PIN
Have	Key, Smart Card
Are	Fingerprint, Face, Iris

Passwords are not reliable.
–Too many
–Can be stolen
–Forgotten

Protect Sensitive Information
–Banking
–Medical

---

Has been used since 14th century in China
–Reliable and trusted

Will never leave at home

Fingerprints are unique
–Everyone is born with one

80% of public has biometric recorded

Has been used since 14th century in China
–Reliable and trusted

Will never leave at home

Fingerprints are unique
–Everyone is born with one

80% of public has biometric recorded

---

6 classes of patterns

---

Minutiae

–Crossover: two ridges cross each other
–Core: center
–Bifurcation: ridge separates
–Ridge ending: end point
–Island: small ridge b/w 2 spaces
–Delta: space between ridges
–Pore: human pore

Two main technologies used to capture image of the fingerprint

–Optical – use light refracted through a prism

–Capacitive-based – detect voltage changes in skin between ridges and valleys

---

Advanced Minutiae Based Algo

Advanced Minutiae Based Algorithm

–Developed by Suprema Solutions

–Two processes

• Feature Extractor
• Matcher

---

Feature Extractor

–Core of fingerprint technology

–Capture and enhance image

–Remove noise by using noise reduction algorithm

–Processes image and determines minutiae

• Most common are ridge endings and points of bifurcation
• 30-60 minutia

–Capture Image

–Enhance Ridge

–Extract Minutiae

–Most frequently used minutiae in applications

• Points of bifurcation
• Ridge endings

–Minutiae Coordinate and Angle are calculated

–Core is used as center of reference (0,0)

---

Matcher

–Used to match fingerprint

–Trade-off between speed and performance

–Group minutiae and categorize by type

• Large number of certain type can result in faster searches

---

Accuracy

–97% will return correct results

–100% deny intruders

Image

–Minutiae is retrieved and template created

• Encrypted data

–Image is discarded

• Cannot reconstruct the fingerprint from data

Several sensors to detect fake fingerprints

–Cannot steal from previous user

• Latent print residue (will be ignored)

–Cannot use cut off finger

• Temperature
• Pulse
• Heartbeat sensors
• Blood flow