Dr. Bernhard Burgeth
(Bld. E2 4 (27.1), room 409, phone 0681-302-64382, burgeth@mia.uni-saarland.de)
Lecture with exercises, 2+1 hours per week, 5 credit points
Time: Friday 11–13 (11 a.m.–1 p.m.),
starting October 21, 2005
Location: Bld. 45, lecture hall 001
Exercises: Bld. 45, lecture hall 001, Mondays,
16:00-18:00 (4 p.m.- 6 p.m.) every second week
Equally suited for students of mathematics, computer science, physics, and civil engineering. Requires undergraduate knowledge in mathematics (e.g. ''Mathematik für Informatiker I–III''). The necessary background knowledge from probability theory, statistics, and information theory will be provided during the lecture.
Pattern analysis/recognition aims at the automatic detection and identification of patterns in data. Pattern means any regularity, relation or structure inherent in data sources. This comprises, for example, 1D signals, any kind of images in digital format, 3D data sets like image sequences, or sequences of 3D images. The detection of patterns is of great importance in many fields connected with computer science.
This lecture introduces the basic goals and techniques of pattern recognition for parametric and non-parametric classification, clustering and feature extraction. Special attention will be paid to the kernel-based methods which embed the data items in a generally high dimensional feature space. In the lecture the construction and the usage of so-called kernel functions in efficient and robust algorithms is discussed.
Language: The lectures will be delivered in English. In the exercise groups both English and German may be used.
Exercises: The next exercise session will take place at Monday, February 6th, 2006.
The egistration period ended on Friday, February 17th.
The oral exams are scheduled for
Thursday 23rd, and Friday 24th of February 2006.
Location : Building E1 1, room 3.07.1
A detailed schedule is now available here . UPDATE (20.02.06, 15:25)
Participants of the course can download the slides here (access password-protected).
| No. | Date | Topic | Exercises |
|---|---|---|---|
| 0 | 21/10 | Introduction | – |
| 1 | 28/10 | First Principles | – |
| 2 | 4/11 | Preliminaries from Stochastics I | – |
| 3 | 11/11 | Preliminaries from Stochastics II | On-Site exercises 14.11.05 |
| 4 | 11, 18/11 | Preliminaries from Stochastics III | Homework included in lecture |
| 5 | 18/11 | Pattern Recognition: First Principles | – |
| 6 | 18, 25/11 | Pattern Recognition: Bayesian Decission Theory | – |
| 7 | 2, 9/12 | Pattern Recognition: MLE and BE (preliminary version!) | On-Site exercises 5.12.05 |
| 8 | 16/12 | Pattern Recognition: PCA and FLD (preliminary version!) | On-Site exercises 19.12.05 |
| 9 | 13/01 | Pattern Recognition: Kernel Methods I | – |
| 10 | 20, 27/01 | Pattern Recognition: Kernel Methods II | On-Site exercises 23.01.06 |
| 11 | 10, 03/02 | Pattern Recognition: Algorithms in Feature Space (update!) | On-Site exercises 06.02.06 |