04-Biometrics-Lecture-4-Part2-2008-10-13

04-Biometrics-Lecture-4-Part2-2008-10-13 - Master SC...

Info iconThis preview shows pages 1–9. Sign up to view the full content.

View Full Document Right Arrow Icon
1 Biometrics http://scgwww.epfl.ch/courses Master SC – Information and Communication Security Dr. Andrzej Drygajlo Speech Processing and Biometrics Group Signal Processing Institute Ecole Polytechnique Fédérale de Lausanne (EPFL) Center for Interdisciplinary Studies in Information Security (ISIS)
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
2 Dynamic Signature Recognition Generalities Feature extraction Signature models and templates Signature recognition errors Signature recognition systems Advantages and disadvantages of signature as biometric
Background image of page 2
3 References S.Y. Kung, M.W. Mak, S.H. Lin, “ Biometric Authentication: A Machine Learning Approach”, Prentice Hall PTR, Upper Saddle River, NJ, 2004 A. Drygajlo, “ Traitement de la parole ”, EPFL, Lausanne, 2003
Background image of page 3

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
4 Models in Signature Recognition Deterministic Methods Dynamic Time Warping ( DTW ) Vector Quantization ( VQ ) Statistical Methods Gaussian Mixture Model ( GMM ) Hidden Markov Model ( HMM )
Background image of page 4
5 Principal structure of signature recognition systems Feature extraction Similarity (Distance) Models for each signature Score Raw data Training Deterministic methods : - Dynamic Time Warping (DTW) - Vector Quantization (VQ) Statistical methods : - Gaussian Mixture Models (GMMs) - Hidden Markov Models (HMMs)
Background image of page 5

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
6 Gaussian Mixture Model (GMM) 12 () (1) ( (2) ) T T T vD v vv v v v v D ⎤⎡ ⎥⎢ ••• •• ⎦⎣ Feature vectors for training GMM Feature 1 Feature 2 Feature D Histograms score = log-likelihood (signature | model)
Background image of page 6
7 Univariate Gaussian distribution -1 0 1 2 3 4 5 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 () 1 x t = ( ( ) 0.242 bxt = 2 2 2 1( ( ) ) (() ) e x p 2 2 xt μ σ πσ =⋅ 2 2 1 = =
Background image of page 7

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
8 Segment models Feature vectors Segment k -1 Segment k Segment k +1 time
Background image of page 8
Image of page 9
This is the end of the preview. Sign up to access the rest of the document.

This note was uploaded on 06/25/2009 for the course MATH MAT 400 taught by Professor Jamespotvein during the Fall '08 term at University of Toronto.

Page1 / 35

04-Biometrics-Lecture-4-Part2-2008-10-13 - Master SC...

This preview shows document pages 1 - 9. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online