CORE ZOO. INTRODUCTION Will biometric characteristics within a zoo menagerie change with the addition of good and bad quality biometric enrollments?

Slides:



Advertisements
Similar presentations
Data Mining in Computer Games By Adib Adam Hussain & Mohammed Sarfraz.
Advertisements

Multiple Analysis of Variance – MANOVA
ECE 5367 – Presentation Prepared by: Adnan Khan Pulin Patel
Chapter 9 Creating and Maintaining Database Presented by Zhiming Liu Instructor: Dr. Bebis.
BIOMETRICS By Lt Cdr V Pravin 05IT6019. BIOMETRICS  Forget passwords...  Forget pin numbers...  Forget all your security concerns...
Creating and Maintaining Databases Dr. Pushkin Kachroo.
C. L. Wilson Manager, Image Group Biometrics Overview of the PATRIOT Act.
Zen and the Art of Facial Image Quality Terry P. Riopka.
The Statistics of Fingerprints A Matching Algorithm to be used in an Investigation into the Reliability of the Use of Fingerprints for Identification Bob.
66: Priyanka J. Sawant 67: Ayesha A. Upadhyay 75: Sumeet Sukthankar.
National Institute of Science & Technology Fingerprint Verification Maheswar Dalai Presented By MHESWAR DALAI Roll No. #CS “Fingerprint Verification.
Biometrics & Security Tutorial 5. 1 (a) Understand two stages (Enrollment and Authentication) in a fingerprint system. (P6: 12)
Cascaded Filtering For Biometric Identification Using Random Projection Atif Iqbal.
Department of Electrical and Computer Engineering Physical Biometrics Matthew Webb ECE 8741.
Comparison and Combination of Ear and Face Images in Appearance-Based Biometrics IEEE Trans on PAMI, VOL. 25, NO.9, 2003 Kyong Chang, Kevin W. Bowyer,
Scatterplots By Wendy Knight. Review of Scatterplots  Scatterplots – Show the relationship between 2 quantitative variables measured on the same individual.
Biometrics Kyle O'Meara April 14, Contents Introduction Specific Types of Biometrics Examples Personal Experience Questions.
TEAM-1 JACKIE ABBAZIO SASHA PEREZ DENISE SILVA ROBERT TESORIERO Face Recognition Systems.
IIIT Hyderabad Atif Iqbal and Anoop Namboodiri Cascaded.
ENTROPY OF FINGERPRINT SENSORS. Do different fingerprint sensors affect the entropy of a fingerprint? RESEARCH QUESTION/HYPOTHESIS.
PSY 307 – Statistics for the Behavioral Sciences
Vision-Based Biometric Authentication System by Padraic o hIarnain Final Year Project Presentation.
Uba Anydiewu, Shane Bilinski, Luis Garcia, Lauren Ragland, Debracca Thornton, Joe Tubesing, Kevin Chan, Steve Elliott, and Ben Petry EXAMINING INTRA-VISIT.
Biometrics: Ear Recognition
CHAMELEON : A Hierarchical Clustering Algorithm Using Dynamic Modeling
Chapter 14: Nonparametric Statistics
BPS - 3rd Ed. Chapter 211 Inference for Regression.
Introduction to Biometrics Charles Tappert Seidenberg School of CSIS, Pace University.
BIOMETRICS. BIOMETRICS BIOMETRICS  Forget passwords...  Forget pin numbers...  Forget all your security concerns...
Secure Systems Research Group - FAU Classifying security patterns E.B.Fernandez, H. Washizaki, N. Yoshioka, A. Kubo.
Symmetric hash functions for fingerprint minutiae S. Tulyakov, V. Chavan and V. Govindaraju Center for Unified Biometrics and Sensors SUNY at Buffalo,
DRS \\ 7jun02 1 Operating Principles for Very small fingerprint sensors.
Biometrics Stephen Schmidt Brian Miller Devin Reid.
Representations for object class recognition David Lowe Department of Computer Science University of British Columbia Vancouver, Canada Sept. 21, 2006.
Patrick Herrmann, Kautilya Madhav, Catherine Muturi, Jack Rosati, Curtis Rose, Jonathan Ruggaard, Ryan Rumble, Kyle Senteney, Ben Petry, Steve Elliott,
Practice Page 65 –2.1 Positive Skew Note Slides online.
Designing multiple biometric systems: Measure of ensemble effectiveness Allen Tang NTUIM.
EXAMINING INTRA-VISIT IRIS STABILITY (VISIT 1) Bo Brown, Jing Guan, Vince Sipocz, Aidan Chamberlain, Brandon Cox, Preston Flint, Eric Hollensbe, Brandon.
Fighting Identity Theft with Advances in Fingerprint Recognition Dick Mathekga.
Introduction to Biometrics Dr. Bhavani Thuraisingham The University of Texas at Dallas Lecture #5 Issues on Designing Biometric Systems September 7, 2005.
8th Grade Forensic Science
Actigraphy Assessment of Mother’s Sleep : 6, 12 & 18 weeks postpartum Introduction: Janelle MacKenzie, Kerry Armstrong & Simon Smith Method: Results: Conclusion:
Introduction to Biometrics Dr. Bhavani Thuraisingham The University of Texas at Dallas Lecture #6 Guest Lecture + Some Topics in Biometrics September 12,
1 Take a challenge with time; never let time idles away aimlessly.
The Practice of Statistics, 5th Edition Starnes, Tabor, Yates, Moore Bedford Freeman Worth Publishers CHAPTER 1 Exploring Data Introduction Data Analysis:
A Closer Look at Fingerprints Image from ftp://sequoyah.nist.gov/pub/nist_internal_reports/ir_6534.pdf T. Trimpe 2007
By Kyle Bickel. Road Map Biometric Authentication Biometric Factors User Authentication Factors Biometric Techniques Conclusion.
BPS - 5th Ed. Chapter 231 Inference for Regression.
ENTROPY OF FINGERPRINT SENSORS STEPHEN ELLIOTT, KEVIN O’CONOOR, ZACH MOORE, JEFF CHUDIK, TORREY HUTCHISON, AND NICK THOMPSON.
1 By maintaining a good heart at every moment, every day is a good day. If we always have good thoughts, then any time, any thing or any location is auspicious.
Tom Face Recognition Software in a border control environment: Non-zero-effort-attacks' effect on False Acceptance Rate.
Practice Page Practice Page Positive Skew.
FACE RECOGNITION TECHNOLOGY
Hu Li Moments for Low Resolution Thermal Face Recognition
Chapter 1: Exploring Data
BIOMETRICS By Lt Cdr V Pravin 05IT6019.
CHAPTER 1 Exploring Data
Good Morning AP Stat! Day #2
Basic Practice of Statistics - 3rd Edition Inference for Regression
Hybrid Finger print recognition
CHAPTER 1 Exploring Data
CHAPTER 1 Exploring Data
Basic Practice of Statistics - 3rd Edition
Basic Practice of Statistics - 3rd Edition
Chapter 1: Exploring Data
CHAPTER 1 Exploring Data
For First Place Most Times Up at the Table
CHAPTER 1 Exploring Data
Chapter 1: Exploring Data
CHAPTER 1 Exploring Data
Presentation transcript:

CORE ZOO

INTRODUCTION

Will biometric characteristics within a zoo menagerie change with the addition of good and bad quality biometric enrollments? RESEARCH QUESTION/HYPOTHESIS

Biometric menageries are used to classify subjects based on biometric matching to their own template and to others There are many different types of zoo’s, like the Doddington zoo and the Dunstone and Yager zoo, and they all differ from each other STATEMENT OF THE PROBLEM

LITERATURE REVIEW

Biometrics refers to recognizing individuals based on behavioral and physical characteristics such as fingerprints and facial features These biometrics can be used to identify and relate one’s features to themselves and to others BIOMETRICS

Fingerprints are more commonly used over every other type of biometric, especially when comparing with zoo menageries Fingerprints have many “minutiae points” across the surface of the finger which are used to compare and identify who a fingerprint belongs to FINGERPRINTS

Quality can be tested by a number of different factors, the fidelity image, the utility image, and the character image Fidelity checks the degree of similarity between the sample and its user Utility assesses the overall impact on the system Character represents features from a biometric sample Higher quality images tend to perform better in a recognition system High quality fingerprints have well-defined ridge structures, good global and local features as well as no abrasions, residues, cuts, or creases from age [4] FINGERPRINT QUALITY

These zoo menageries are created by matching a person’s biometric sample against a population. If they are not the same person’s fingerprint it is given an imposter or genuine score that describes them as imposters or genuine These scores determine your zoo classification IMPOSTER AND GENUINE SCORES

Zoo menageries were created because there are still flaws within biometrics They characterize people as different animals depending on how well their fingerprint compares to another There are two main types of zoo menageries, the Doddington zoo and the Dunstone Yeager zoo ZOO MENAGERIES

Measures individual scores based on how well a user matches to those in a specific dataset [8] Characterizes people as sheep, wolves, goats, or lambs Sheep match well with their own template Wolves match well with others Goats don’t match themselves well Lambs match with other people but not themselves DODDINGTON ZOO

The Dunstone and Yager Zoo is created by matching a person’s biometric sample against the data set Measures the relationship between a user’s genuine and impostor match results [5] Characterizes people as worms, phantoms, chameleons, and doves Worms match to others well but not to themselves Phantoms don’t match well with anybody Chameleons match to themselves and others well Doves match to themselves well but not to others DUNSTONE AND YAGER ZOO

The subject is given an average genuine based on their match results to themselves and an impostor match score based on their match results to others These scores determine what zoo classification each subject belongs in GENUINE/IMPOSTOR SCORES

Phantoms WormsChameleons Doves

Davis Wittman did a study that shows that zoo animals do exist in biometric measurements Found outliers in facial recognition that closely resemble other people as well as ones that do not match well to themselves DOES A ZOO EXIST?

Wittman et al [6] did a study that shows that zoo animals do exist in biometric measurements Found outliers in facial recognition that closely resemble other people as well as ones that do not match well to themselves WEAKNESS OF THE ZOO

DOVES GHGH GLGL IHIH ILIL Top 25% Bottom 25% Genuine Impostor

WORMS GHGH GLGL IHIH ILIL Top 25% Bottom 25% Genuine Impostor

CHAMELEONS GHGH GLGL IHIH ILIL Top 25% Bottom 25% Genuine Imposture

PHANTOMS GHGH GLGL IHIH ILIL Top 25% Bottom 25% Genuine Imposture

METHODOLOGY

Core – D1 – Original dataset from O’Connor’s thesis Dataset 2 – D2 – additional data set from Benny Dataset 3 – D3 – additional data set from DHS Top- Top 25% of images in terms of quality Bottom – Bottom 25% of images in terms of quality CLASSIFICATIONS OF DATA SETS

The datasets were combined into groups as shown below: 1. Core plus dataset 2 top as well as core plus dataset 2 bottom (C+D2T and C+D2B) Results are in one table 2. Core plus dataset 3 top as well as core plus dataset 3 bottom (C+D3T and C+D3B) Results are in one table 3. Core plus dataset top (C+D2T+D3T) 4. Core plus dataset bottom (C+D2B+D3B) Results for 3 and 4 are in one table METHODOLOGY

HARDWARE USED Crossmatch Guardian Fingerprint sensor

Filemaker Database of Samples OWR Bio-Metrics Used to check zoo placement Megamatcher Fingerprint matching and quality scores SOFTWARE USED

RESULTS

154 subjects in Core 17 subjects added from Dataset 2 for top and bottom 20 subjects added from Dataset 3 for top and bottom RESULTS

Core + Bottom 25%CoreCore + Top 25% IDClassificationIDClassificationIDClassification 358Normal358Phantoms358Phantoms 652Normal652Phantoms652Normal 677Normal677Phantoms677Normal 697Normal697Phantoms697Normal 704Normal704Doves704Normal 721Normal721Doves721Normal 724Chameleons724Chameleons724Normal 741Normal741Normal741Worms 742Normal742Phantoms742Phantoms 743Normal743Normal743Chameleons 747Normal747Phantoms747Normal 839Normal839Phantoms839Normal RESULTS OF CORE + D2

Core + Bottom 25%CoreCore + Top 25% IDCategoryIDCategoryIDCategory 239Worm239Normal239Normal 302Dove302Dove302Normal 358Normal358Phantom358Phantom 359Chameleon359Chameleon359Normal 724Chameleon724Chameleon724Normal 726Chameleon726Chameleon726Normal 737Worm737Worm737Normal 740Phantom740Normal740Phantom 743Chameleon743Normal743Normal 775Normal775Normal775Phantom 839Normal839Phantom839Phantom RESULTS OF CORE + D3

RESULTS OF CORE + D2 + D3 Core + Bottom 25%CoreCore + Top 25% IDCategoryIDCategoryIDCategory 302Dove302Dove302Normal 358Normal358Phantom358Phantom 359Chameleon359Chameleon359Normal 677Phantom677Phantom677Normal 686Phantom686Phantom686Normal 697Phantom697Phantom697Normal 704Dove704Dove704Normal 721Dove721Dove721Normal 724Chameleon724Chameleon724Normal 726Chameleon726Chameleon726Normal 740Phantom740Normal740Normal 741Normal741Normal741Worm 743Chameleon743Normal743Chameleon 747Phantom747Phantom747Normal 839Normal839Phantom839Normal

EXAMPLE SUBJECT 839

SUBJECT 839 CORE GROUP

SUBJECT D2 TOP

SUBJECT D2 BOTTOM

SUBJECT D3 TOP

SUBJECT D3 BOTTOM

SUBJECT D2 + D3 TOP

SUBJECT D2 + D3 BOTTOM

Low quality images tend to add subjects with low genuine scores and low imposter scores High quality images tend to add subjects with average to high genuine scores and low imposter scores Low quality additions also shifted the entire plot lower on the genuine axis OBSERVATIONS

CONCLUSIONS

Image quality does have some effect on the zoo characteristics of fingerprints Up to 10% of subjects changed zoo animals Even though the impostor scores have minimal change, subjects are still changing animal classifications CONCLUSIONS

Use larger and more variable populations such as top and bottom 25% together Are subjects stable when adding animal classifications from other datasets? Examine with fat plots instead of categorical analysis Can systems combine and still have the same results? – Interoperability of the zoo FUTURE WORK

REFERENCES