UnivIS
Informationssystem der Friedrich-Alexander-Universität Erlangen-Nürnberg © Config eG 

Pattern Recognition (PR)

Dozent/in
Dr.-Ing. Dieter Hahn

Angaben
Vorlesung
3 SWS, Schein, ECTS-Studium, ECTS-Credits: 5
geeignet als Schlüsselqualifikation
Zeit und Ort: Mo 12:00 - 14:00, H10; Mi 13:30 - 14:30, H10
ab 20.10.2010

Studienfächer / Studienrichtungen
WPF CE-MA 5-7
WPF INF-DH-ME 5-7
WPF INF-DH-MI 5-7
WPF IuK-DH-MMS-INF1 5-7

ECTS-Informationen:
Title:
Pattern Recognition

Credits: 5

Contents
This lecture gives an introduction into the basic and commonly used classification concepts. First the necessary statistical concepts are revised and the Bayes classifier introduced. Further concepts include generative and discriminative models like logistic regression, the Gaussian classifier, Linear Discriminant Analysis, the Perceptron and Support Vector Machines (SVMs). Finally more complex methods like the Expectation Maximization Algorithm and Hidden Markov Models are discussed.
In addition to the mentioned classifiers, methods necessary for practical application like dimensionality reduction, optimization methods and the use of kernel functions are explained.
In the tutorials the methods and procedures which are presented in this lecture are illustrated using theoretical and practical exercises.

Literature
  • lecture notes

  • Duda R., Hart P. and Stork D.: Pattern Classification

  • Niemann H.: Klassifikation von Mustern

  • Niemann H.: Pattern Analysis and Understanding

  • Fu K.S.: Sequential Methods in Pattern Recognition and Machine Learning

  • Schürmann J.: Polynomklassifikatoren für die Zeichenerkennung

Zusätzliche Informationen
Schlagwörter: Mustererkennung, Vorverarbeitung, Merkmale, Klassifkation
Erwartete Teilnehmerzahl: 30
www: http://www5.informatik.uni-erlangen.de/lectures/ws-1011/pattern-recognition-pr/

Zugeordnete Lehrveranstaltungen
UE: Pattern Recognition Exercises
www: http://www5.informatik.uni-erlangen.de/lectures/ws-1011/pattern-recognition-pr/

Verwendung in folgenden UnivIS-Modulen
Startsemester WS 2010/2011:
Pattern Recognition (PR)

Institution: Lehrstuhl für Informatik 5 (Mustererkennung)
UnivIS ist ein Produkt der Config eG, Buckenhof