back to Biometrics (CSE 40537/60537)

Progress

_______________________________________
Week 1

Tue. 1/16: Introduction and syllabus
Discussed in class:
Organization of the course. Definition of biometrics and recognition types that can be accomplished by a biometric system.
Slides:
Syllabus (all), What is biometrics? (1->9)

Thu. 1/18: What is Biometrics?
Discussed in class:
Basic biometric vocabulary, biometric modes and their expected properties.
Slides:
What is biometrics? (10->25)
Reading materials:
Jain et al., "Introduction to Biometrics":
pp. 1-24 (including Sec. 1.4.1.1 "Verification system error rates")

_______________________________________
Week 2

Tue. 1/23: What is Biometrics?
Discussed in class:
Enrollment vs authentication, laboratory setups vs large-scale systems, basic biometric errors: FMR, FNMR, EER, FTA and FTE.
Slides:
What is biometrics? (26->end)

Thu. 1/25: Fingerprints recognition
Discussed in class:
Fingerprint ridges and valeys, level 1 features: singular points, detection of singular points, classification of fingerprints
Slides:
Fingerprints recognition (1->22)
Reading materials:
Maltoni et al., "Handbook of Fingerprint Recognition":
a) pp. 34-41: Formation of fingerprints (Sec. 1.9) -> Fingerprint Representation and Feature Extraction (Sec. 1.12)
b) pp. 120-130: Singularity and Core Detection (Sec. 3.5)

_______________________________________
Week 3

Tue. 1/30: Fingerprints recognition
Discussed in class:
Level 2 features: minutiae. Level 3 features: pores, ridge shape, warts, scars, etc. Acquisition of fingerprint samples: capacitive and optical sensors.
Slides:
Fingerprints recognition (23->37)
Reading materials:
Maltoni et al., "Handbook of Fingerprint Recognition":
pp. 63-71: Live-Scan Fingerprint Sensing (Sec. 2.3)

Thu. 2/1: Fingerprints recognition
Discussed in class:
Acquisition of fingerprint samples: piezoelectric, ultrasound and contactless sensors. Fingerprint image enhacement using wavelet expansion.
Slides:
Fingerprints recognition (38->59)
Reading materials:
Maltoni et al., "Handbook of Fingerprint Recognition":
pp. 131-141: Enhancement (Sec. 3.6, excluding Sec. 3.6.4)

_______________________________________
Week 4

Tue. 2/6: Biometric data acquisition: fingerprints

Thu. 2/8: Fingerprints recognition
Discussed in class:
Fingerprint image segmentation, detection of minutiae, removing "fake" minutiae, Hough transform for image alignment, minutiae matching.
Slides:
Fingerprints recognition (60->end)
Reading material:
Maltoni et al., "Handbook of Fingerprint Recognition":
a) pp. 143-157: Minutiae Detection (Sec. 3.7)
b) pp. 177-194: Fingerprint Matching, Minutiae-Based Methods (Sec. 4.3)
Optional materials on Gabor filtering:
a) Daugman, "Computer Vision Notes": pp. 36-40 (Sec. 6.2 - 6.4)
b) 2D Gabor filtering in Matlab (try it out!)

_______________________________________
Week 5

Tue. 2/13: Face recognition
Discussed in class:
Importance of face recognition, face detection, region-based and sliding-window-based methods, Viola-Jones algorithm.
Slides:
Face recognition (1->22)
Reading material:
Daugman, "Computer Vision Notes": pp. 80-92 (Sec. 16)

Thu. 2/15: Face recognition
Discussed in class:
Face alignment (pose, illumination), feature extraction and matching (LBP and CNN).
Slides:
Face recognition (23->end)
Additional reading materials:
a) "Convolutional Networks", Chapter 9 in: Goodfellow, Bengio, Courville, "Deep learning"
b) YT videos in slides: 3D face alignment, LAX using facial recognition for British Airways boarding

_______________________________________
Week 6

Tue. 2/20: Biometric data acquisition: face

Thu. 2/22: Iris recognition
Discussed in class:
Iris formation, verification of well promulgated statements: 1) genetic independence, 2) aging, 3) post-mortem iris recognition. Visible-light vs near-infrared-ligh acquisition, basic properties of an iris image.
Slides:
Iris recognition (1-36)
Reading material:
a) J. Daugman, "How Iris Recognition Works", IEEE CSVT, 14(1), 2004
b) Interesting summary of Daugman's IAPR Award lecture

_______________________________________
Week 7

Tue. 2/27: Iris recognition
Discussed in class: iris image acquisition (cont'd), iris image ISO formats, Daugman's iris segmentation algorithm (integro-differential operator).
Slides: Iris recognition (37-58)

Thu. 3/1: Iris recognition
Discussed in class: Alternative segmentation methods (Fourier expansion, active contours, deep-learning-based solutions), iris image normalization and filtering
Slides: Iris recognition (59-80)

_______________________________________
Week 8

Tue. 3/6: Iris recognition
Discussed in class: Daugman's coding and matching, statistical properties of an iris code, masking of iris code bits (occlusions, "fragile bits").
Slides: Iris recognition (81-end)

Thu. 3/8: Hand biometrics (silhouette, palmprint, veins)
Discussed in class: What kind of hand features we can use in biometrics? Hand geometry, feature extraction, feature classification. 
Slides: Use of hand in biometrics (1-27)

_______________________________________
Midterm break (3/10 - 3/18)

_______________________________________
Week 9

Tue. 3/20: Biometric data acquisition: iris

Thu. 3/22: Hand biometrics (silhouette, palmprint, veins)
Discussed in class: Palm vein and finger vein biometrics. Briefly: use of hand thermal maps in identification and liveness detection.
Slides: Use of hand in biometrics (28-end)

_______________________________________
Week 10

Tue. 3/27: Handwritten signatures
Discussed in class: Off-line and on-line signatures, signature components, data acquisition.
Slides: Recognition of handwritten signatures (1->19)
Reading material:
J. Putz-Leszczynska, Signature Verification -- A Comprehensive Study Of The Hidden Signature Method, Int. J. Appl. Math. Comput. Sci., 2015, Vol. 25, No. 3, pp. 659-674

Thu. 3/29: Handwritten signatures
Discussed in class: Global and local features, feature classification. We have started a discussion what the dynamic time warping is.
Slides: Recognition of handwritten signatures (20->41)

_______________________________________
Week 11

Tue. 4/3: Biometric data acquisition: signatures

Thu. 4/5: Handwritten signatures
Discussed in class: Dynamic Time Warping
Slides: Recognition of handwritten signatures (42->end)

_______________________________________
Week 12

Tue. 4/10: Speaker recognition
Discussed in class: Speaker vs speech recognition, formants (Lucier's experiment), voice features in time domain (auto-regressive models, ICA) and frequency domain. "Mel" scale.
Slides: Speaker recognition (1->28)

Thu. 4/12: Speaker recognition (cont'd). Security: system vulnerabilities and presentation attacks
Discussed in class: Cepstral features in speaker recognition. Biometric system and its vulnerabilities, presentation attacks in iris, fingerprint and vein recognition systems.
Slides: Speaker recognition (29->end), Security of biometrics (1->34)

_______________________________________
Week 13

Tue. 4/17: Biometric data acquisition: gummy fingers and printed irises
Prerequisities: see the instructions how to prepare gummy fingers and iris printouts at home

Thu. 4/19: Security: presentation attack detection (in fingerprint, iris and face biometrics)
Discussed in class: Definition of PAD, subversive (what we have to detect) vs suspicious (what we can detect) actions, use of static and dynamics properties of a finger for PAD
Slides: Security of biometrics (35->55)
Reading material: Fingerprint PAD survey

_______________________________________
Week 14

Tue. 4/24: Security: presentation attack detection (in fingerprint, iris and face biometrics)
Discussed in class: Detection of sweat pores (as the fingerprint PAD), detection of "fake frequencies" (as the iris PAD).
Slides: Security of biometrics (56->76)
Reading material: Iris PAD survey, Face PAD survey

Thu. 4/26: Biometric system's security
Discussed in class: Pupil dynamics for iris PAD. Hill climbing attacks, "master" fingerprints, use of synthetic biometric data, biometric cryptography for cancelable biometrics.
Slides: Security of biometrics (77->110)

_______________________________________
Week 15

Tue. 5/1: Biometric system's security, performance evaluation of biometric systems
Discussed in class: Match-off-card and match-on-card biometric systems. Technology, scenario and operational evaluations, ROC, DET, FTA, FTE.
Slides: Security of biometrics (111->end), Statistical evaluation of biometrics (1-32)

_______________________________________
Final test

Monday, May 7, 10:30am - 12:30pm, room: TDB