Symbolbild
Ein Symbolbild, das die AG repräsentiert
Login |
 
 

Mustererkennung

Lecturer and Tutor

Prof. Dr. Dietmar Saupe (lecturer)

Igor Zingman (tutor)

Schedule

Lectures Mo. 10:00 - 11:30 Z613
  We. 10:00 - 11:30 Z613
Lab Session Th. 15:15 - 16:45 Z613
Exam Tu. 10:00 - 11:45 A704

Note that the last lab session will be held as usual on Thursday, 24.07

Target Audience and Prerequisites

The course belongs to the topic areas Foundations of Computer Science and Applied Computer Science and is for Master and PhD-Students (not Bachelor) of the following subjects:

  • Information Engineering and Computer Science
  • Mathematics
  • Physics

Basic knowledge in probability theory, calculus and linear algebra are expected.

Content

This one-semester graduate course provides a first introduction to statistical inference - theory that lies in the heart of pattern recognition (classification). The course includes the following topics of classical theory:

  • Bayesian inference,
  • parametric and nonparametric estimation of probability density functions,
  • linear classifiers,
  • classifiers evaluation,
  • feature generation and feature selection,

complemented by discussions of newer approaches, such as, support vector machines or neural networks.

Course works and credits

There are 2 lectures + 1 lab a week. At the lab sessions, students will work on theoretical and computer-based problems that constitute a part of homework assignments.

The marks will affect your final grade as follows:

  • > 50 % admission for final exam
  • > 70 % 1/3 of a grade bonus on your final grade
  • > 90 % 2/3 of a grade bonus on your final grade

Lab Sheets

Lab Sheet Data Deadline

pat01.pdf

pattern1.mat no submission

pat02.pdf

  15.05
 pat03.pdf   22.05
pat04.pdf   02.06
pat05 (comp ex.)  data 18.06
pat06   04.07
pat07   24.07

Matlab Access

Matlab will be used extensively during the course and is installed locally on every computer in Z 613. For completing assignments you can use the computer pool in V 304. Remote access from home might be slow. Matlab is installed on a server:

  • for Linux: /net/lin_local/matlab/bin/matlab
  • for Windows: \\titan01\lin_local\matlab\win\bin\matlab.bat

 

Matlab tutorial

Exam

The course ends with a written final exam. The final grade will be the grade obtained at the final exam.

 

The exam is held on Tu 29.07, 10:00-11:45 in room A704.

Literatur / Weblinks

The main textbook

  • Richard O. Duda, Peter E. Hart and David G. Stork.
    Pattern Classification
    (2nd ed.), 2000

Complementary sources

  • Slides from a course at the University of Western Ontario. 2004, 2006.
  • Sergios Theodoridis, Konstantinos Koutroumbas. (feature selection, Ch. 5)
    Pattern Recognition
    (3d edition), 2006.

 

Additional literature

  • Christopher M. Bishop.
    Pattern Recognition and Maschine Learning
    , 2007.
  • Keinosuke Fukunaga.
    Introduction to Statistical Pattern Recognition
    (2nd edition), 1990.