biometrics: machines recognizing people - carleton · biometric template security it is claimed to...
TRANSCRIPT
![Page 1: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/1.jpg)
Biometrics: Machines recognizing people
Biometrics & Authentication
Technologies: security issues
Andy Adler
Systems and Computer Engineering, Carleton
![Page 2: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/2.jpg)
Finger anatomy
![Page 3: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/3.jpg)
Fingerprint: Rolled ink
Ink Roller
![Page 4: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/4.jpg)
Fingerprints: Optical
Scanner
![Page 5: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/5.jpg)
Fingerprints: Capacitive scanner
![Page 6: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/6.jpg)
Cleaned fingerprint
![Page 7: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/7.jpg)
Get features: minutiae
![Page 8: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/8.jpg)
Fingerprint: Compare
Optical Scanner
1998
CapacitiveScanner
2004
![Page 9: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/9.jpg)
Get features: minutiae1998
2004
Compareand
Decide
![Page 10: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/10.jpg)
Fingerprint examples
Thumbs from my family
Age 35 Age 65Age 34Age 6Age 4
![Page 11: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/11.jpg)
Are fingerprints unique?
What do you mean by unique?
Real Question:
Are fingerprints distinguishable?
Cut
Moved
Lesspressure
![Page 12: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/12.jpg)
What does Unique mean?
� No differences at all
� But then fingers change every day
� Detectably different
� But our detection algorithm keep getting better
� How informative is a fingerprint
� “the decrease in uncertainty about the identity
from a biometric measurement”
![Page 13: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/13.jpg)
Face Recognition:
same person?
![Page 14: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/14.jpg)
Same
person?
� I have just demonstrated a massively parallel face recognition computer
� Question:Are computers better or worse than people at faces?
Yes
![Page 15: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/15.jpg)
How do computers recognize
faces?
EigenfacesLandmarks
![Page 16: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/16.jpg)
Today’s FR algs are better than
half of people
![Page 17: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/17.jpg)
Results
� Error rates are high
� Significant improvement in SW 1999-2006
� Most recent algs outperform about half of
people
� No significant difference male/female
![Page 18: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/18.jpg)
Iris
![Page 19: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/19.jpg)
Iris: Processing
RemoveObscuredImage parts
![Page 20: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/20.jpg)
Template DB
How is this used?
Present Template
BiometricCompare
MatchScore
TemplateEnroll Feature
Extraction
FeatureExtraction
Threshold
Decision
![Page 21: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/21.jpg)
What can go wrong?
1-5%1-5%5%False non-match
10ppm10ppm1%False Match
10%10% 3%Failure to
acquire
7% 3%0%Failure to enroll
IrisFingerFace
Very approximate values! Depends on all sorts of things
![Page 22: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/22.jpg)
Biometrics Vulnerabilities
Taxonomy (from Maltoni et al, 2003):
� Circumvension
� Covert acquisition
� Collusion / Coercion
� Denial of Service
![Page 23: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/23.jpg)
Biometrics Security Issues
� Biometrics are not secrets
� Biometrics cannot be revoked
� Biometrics have secondary uses
![Page 24: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/24.jpg)
Template DB
What else can go wrong?
Biometric“Live Image”
Template
BiometricCompare
MatchScore
Template
Biometric“Enrolled
Image”
FeatureExtraction
FeatureExtraction
Threshold
Decision
Regenerateimage
Replay at sensor
Databaseintegrity
Replay at matcher
ID cardintegrity
Modifythreshold
Reverseengineeralgorithm
Modifyweightingof fusion
Enrolldifficult image
Lookalike
Impostor
FraudulentEnrolment
ConfuseAlgorithm(tilt head,squash finger)�
Fatiguedoperator
Invaliddata
Multipleenrolment
![Page 25: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/25.jpg)
What else can go wrong?
Biometric system
Identity
verificationsystem
Release
Cryptokeys
Single
Sign-on
sub-Lookout
system
Authenticate
Credit cardAuthenticateInternet app
Supervised
sensor
unsupervised
desktop
Authenticatevia internet
unsupervisedpublic
![Page 26: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/26.jpg)
Spoofing
![Page 27: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/27.jpg)
Who manages registration?
User locks phone with fingerprint
Cell PhoneIndividual
Voiceprint. Callback
to validate sales
Credit cardIndustry
Iris for fast passenger
processing
PassportGovernment
ExampleWhatWho
![Page 28: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/28.jpg)
Spoofing
Registration fraud
Theft and modification
Duplication
Theft
Privacy worries
Secondary use of data
“Dumpster Diving”
Phishing
Cell
phone
Credit
Card
Pass-port
Vulnerable
Secure
11112
11112
11112
11112
11112
11112
![Page 29: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/29.jpg)
More details / my research …
Biometrics Security
� Biometric uniqueness / entropy
� Biometric template protection
� Flaws in biometric encryption
![Page 30: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/30.jpg)
information content of a
biometric measurement?
Or
� How much do we learn (about identity) from a biometric image
Or
� How much privacy do we loose on releasing a biometric image
![Page 31: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/31.jpg)
Example: measure Height� Measure #1 (at doctor’s office, ie. accurate)
� Measure #2 (via telescope, ie. inaccuate)
Overall
Distribution
FeatureVariability
(high heels,
carry backpack)
MeasurementVariability
(device errors)
![Page 32: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/32.jpg)
Example: measure Height
� How much information learned?
Measure #2
Measure #1
LowAlmost zero
Quite a lotLow
Tall
(7½’ tall)
Average
(5½’ tall)
Know aboutHuman heights Measure
Know about:Human heights
Person’s height
![Page 33: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/33.jpg)
Proposed measure:
relative entropy D(p||q)� Given biometric feature vector x
� Distributions
� intra-person distribution, p(x)
� inter-person distribution, q(x)
� D(p||q) measures inefficiency of assuming q when true distribution is p
Or,
� D(p||q) measures extra information in p than q
![Page 34: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/34.jpg)
Applications: biometric
� Meta algorithm
� Evaluate a new biometric feature
� Biometric Performance limits
� Template size limits
� Inherent match performance limits
� Feasibility of Biometric Encryption
� Limits to Key Length
![Page 35: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/35.jpg)
Applications: abstract
� Quantify privacy
� What is the privacy risk due to the release of certain information?
� What is the privacy gain in obscuring faces?
� Uniqueness of biometrics
� Approach to address: “Are faces / fingerprints / irises unique?”
![Page 36: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/36.jpg)
Biometric template security
It is claimed to be impossible or infeasible to recreate the enrolled image from a template.
Reasons:
� templates record features (such as fingerprint minutiae) and not image primitives
� templates are typically calculated using only a small
portion of the image
� templates are much smaller than the image
� proprietary nature of the storage format makes templates
infeasible to "hack".
![Page 37: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/37.jpg)
Images can be regenerated…?� Typical Biometric processing
� Question: Is this possible?
enrolled“Image”
Template
BiometricCompare
MatchScore
Template
regenerated“Image”
live“Image”
![Page 38: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/38.jpg)
A
B
Iteration
4000
Target
Image
Iteration
600
Iteration
200
Initial
Image
Hill-climbing: begin with a guess, make small modifications; keep modifications which increase the match score
Results:
![Page 39: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/39.jpg)
Improved regenerated image
Average of 10
Best EstimatesTarget Image
• Recently, this approach has been extended to
fingerprint images (Uludag, Ross, Capelli)
![Page 40: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/40.jpg)
Implications: image
regeneration1. Privacy Implications
� ICAO passport spec. has templates encoded with public keys in contactless chip
� ILO seafarer’s ID has fingerprint template in 2D barcode on document
![Page 41: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/41.jpg)
Implications: image
regeneration2. Reverse engineer algorithm
� Regenerated images tell you what the algorithm
‘really’ considers important
Alg. #3Alg. #2Alg. #1Target doesn’t care
about nose
width
![Page 42: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/42.jpg)
Implications: image
regeneration
3. Crack biometric encryption
Biometric encryption seeks to embed a key into the template. Only a valid image will decrypt the key
� Since images varyEnrolled image + Δ => release key
� HoweverEnrolled image + Δ + ε => no release
If we can get a measure of how close we are, they we can get a match score
![Page 43: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/43.jpg)
Biometric Encryption� Recent paper by Ontario Information and
Privacy Commissioner
� “Biometric Encryption: A Positive-Sum Technology that Achieves Strong Authentication, Security AND Privacy”
� A. Cavoukian, A. Stoianov
My concern:
� Biometric Encryption (and biometric cryptographic schemes in general) only offer benefits if they are cryptographically secure.
![Page 44: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/44.jpg)
From: http://www.ipc.on.ca/images/Resources/up-1bio_encryp.pdf
![Page 45: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/45.jpg)
Crack biometric encryption
� Construct match-score from number of
matching elements in link table
� Use quantized template reconstructor
enrolled
Perc
ent
matc
hed
iteration
![Page 46: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/46.jpg)
Fuzzy Vaults for fingerprints (Clancy, 2003)�
Raw Fingerprint With minutiae With added “chaff”
![Page 47: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/47.jpg)
Collusion Attack� Users’ fingerprints may be associated with
many vaults.
� Ex: In the smart card implementation, users will likely carry multiple smart cards associated with different companies, each locked with the same
fingerprint.
� Fuzzy Vault is insecure when the same
fingerprint is used to lock multiple vaults
![Page 48: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/48.jpg)
Biometrics in Canada (Gov't)�
� Passports
� Immigration
� Customs
� Defence
� Natural Resources
� Public Safety
� RCMP
![Page 49: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/49.jpg)
Epilogue: Our future?
Operator: "Thank you for calling Pizza Hut."
Customer: “One All-Meat Special..."
Operator: "Thank you, Sir. Your voice print verifies with your National ID Number: 6102049998"
Customer: (Sighs) "I'd like to order an All-Meat Special pizza..."
Operator: "I don't think that's a good idea, sir."
Customer: "Whaddya mean?"
Operator: "Sir, your medical records indicate that you've got very high blood pressure and cholesterol. Your Health Care provider won't allow such an unhealthy choice."
Customer: "Darn. What do you recommend, then?"
![Page 50: Biometrics: Machines recognizing people - Carleton · Biometric template security It is claimed to be impossible or infeasible to recreate the enrolled image from a template. Reasons:](https://reader033.vdocuments.mx/reader033/viewer/2022060401/5f0e2ff57e708231d43e0730/html5/thumbnails/50.jpg)
Epilogue:
Operator: "You might try our low-fat Soybean Yogurt Pizza. I'm sure you'll like it"
Customer: "What makes you think I'd like something like that?"
Operator: "Well, you checked out 'Gourmet Soybean Recipes' from your local library last week, sir."
Customer: “OK, lemme give you my credit card number."
Operator: "I'm sorry sir, but I'm afraid you'll have to pay in cash. Your credit card balance is over its limit."
Customer: "@#%/$@&?#!"
Operator: "I'd advise watching your language, sir. You've already got a July 2012 conviction for cussing … "