CS533/CmpE533
Pattern Recognition - Spring 2007-08

Class Meeting Times
Mon 8:40 am to 9:55 am
Wed 8:40 am to 9:55 am

   
     
Intructor: Sohaib Ahmad Khan
sohaib at lums dot edu dot pk
http://web.lums.edu.pk/~sohaib


TA: Syed Farooq Ali
farooqali at lums dot edu dot pk
  Instructor Office hrs: (Rm 407 Library Bldg)
Mon 10:15 am - 11:45 am
Wed 10:15 am - 11:45 am


TA Office hrs: (Rm 437)
TBA

Course Description: This course provides an introduction to the area of Statistical Pattern Recognition. The course will be beneficial to graduate students intending to pursue research in this area, as well as in applied fields which use pattern recognition, such as speech recognition, computer vision, image processing, signal classification, optical character recognition and data mining. Major topics covered in the course include supervised and unsupervised learning, Bayesian decision theory, parametric and non-parametric density estimation methods, linear discriminant functions and clustering methods.

Text Book
Pattern Classification (2nd Ed.), Richard O. Duda, Peter E. Hart, David G. Stork, Wiley-Interscience 2001

Reference Books
Statistical Pattern Recognition, 2nd Ed, Andrew Webb, Wiley 2002
Pattern Recognition and Machine Learning, Christopher Bishop, Springer 2006

Course Outline [PDF]

Useful Links
Website of Duda/Hart/Stork Textbook
Websites for reference texts: Webb, Bishop
The Pattern Recognition Files
A good set of lecture slides: PRISM lectures
PR description and links on AAAI website
A good review paper of statistical pattern recognition by Jain et. al published in 2000.
A good glossary of statistical pattern recognition terms by Thomas Minka


Date Lecture Topics Readings Annoucements
Mar 17, 2008 Lecture 1: Course Introduction and Policies. Introduction to Statistical Pattern Recognition. Features and Classifiers. Supervised, Unsupervised and Reinforcement Learning. Generative vs Discriminative Approaches. Estimation Error. Curse of Dimensionality
Ch 1 of DHS Homework 0 assigned
Mar 19, 2008 Lecture 2: Pattern Recognition design cycle, Minimum error rate classification for two category data, Bayes decision rule for two categories

Ch 1 and first four pages of Ch 2 of DHS

Viewgraphs from lecture

 
Mar 24, 2008 Lecture 3: Bayes Decision Theory for Continuous Features, Conditional Risk, Bayes Decision Rule for general case Ch 2 of DHS  
Mar 26, 2008 Lecture 4: Bayes Decision Rule (Two Category Case), Min Error Rate Classification, Minimax Criterion, Discriminant Functions

Sections 2.2.1, 2.3, 2.4 of DHS

Viewgraphs from the lecture

IMPORTANT:
Homework 1: DHS Ch2, Problems: 1, 2, 3, 4, 8, 9, 10, 12, 13, 14
Due Date: Thu Apr 3 at 5:00 pm
Mar 31, 2008 Lecture 5: The Multivariate Normal Density, Eigen Vectors of the Covariance Matrix, Whitening Transformation Section 2.5 of DHS  
Apr 2, 2008 Lecture 6: Simultaneous Diagonalization for Two Normal Densities, Discriminant Functions of the Normal Density Section 2.6  
Apr 7, 2008 Lecture 7: Discriminant Functions of the Normal Density (continued), Error Probabilities and Integrals, Chernoff Bound, Bhattacharyya Bound Section 2.6, 2.7  
Apr 9, 2008 Lecture 8: Parametric Density Estimation, MLE, MLE of mean and variance of Gaussian Distribution Sec 3.1, 3.2  
Apr 14, 2008 Lecture 9: Methods of Evaluating Estimators Sec 3.2.4 and class notes  
Apr 16, 2008 Lecture 10: Examples of computing Bias and MSE of estimators, Review for Midterm

Class notes

Viewgraphs of first 10 lectures

Practice Midterm 1
Practice Midterm 2

IMPORTANT:
Homework 2: [pdf]

Due Date: Mon Apr 28 at 5:00 pm
Apr 21, 2008
Midterm Exam
8:35 am in SC3
Apr 23, 2008 Discussion on Midterm    
Apr 28, 2008 Bayesian Estimation    
Apr 30, 2008 Dimensionality Reduction: Principal Component Analysis   IMPORTANT: Homework 3 (Download handout and data)