IUPR Research Group (DFKI and UniKL)
 
Search Site

Upcoming Courses

  • Summer 2009
    • Bachelor
      • [SEMINAR] Grundlagen der Mustererkennung
    • Bachelor / Master
      • Introduction to Image Processing and Image Understanding
      • Document and Content Analysis
    • Master
      • [LAB] Intelligent Data Mining and Pattern Recognition Competition
  • Winter 2009/2010
    • Bachelor
      • [SEMINAR] Grundlagen der Mustererkennung
    • Bachelor / Master
      • Introduction to Pattern Recognition
      • Privacy, Identity, and Computational Forensics
    • Master
      • [LAB] Computer Vision and Image Understanding
  • Summer 2010
    • Bachelor
    • Bachelor / Master
      • Introduction to Image Processing and Image Understanding
      • Tentatively Offered: Bayesian Decision Theory and Statistical Learning
    • Master
      • [LAB] Intelligent Data Mining and Pattern Recognition Competition
See below for a description of these courses.

Please also see additional Planned Courses.

Intermediate and Advanced Courses


Introduction to Image Processing and Image Understanding

Usually offered every summer.

Hours: 2+1
Textbook: Gonzales and Woods (optional)
Course Language: English (lectures), English or German (exams, exercises)

The course will familiarize students with the basic theory and implementation of image processing and image understanding software through a series of sample problems and projects.The course covers roughly:

  • image types and representations
  • models of image formation
  • color, color spaces, color perception
  • linear filters
  • mathematical morphology
  • edge, corner, and line detection
  • wavelets
  • image and video compression
  • noise models and detection theory
  • shape matching and object recognition
  • motion and stereo image analysis
Algorithms will be illustrated with a variety of applications, including photographic imaging, color correction, image enhancement, industrial inspection, document image analysis and OCR, medical imaging, and forensics.  The course is accompanied by practical exercises in Numerical Python.

Introduction to Pattern Recognition

Usually offered every winter.

This course replaces "Pattern Recognition and Statistical Learning".

Hours: 2+1
Textbook: Duda, Hart, and Stork (optional)
Course Language: English (lectures), English or German (exams, exercises)

The course introduces students to the most common algorithms in pattern recognition and statistical learning:

  • generative data models and model-based classification
    • black box parameter recognition
  • nearest neighbor classification
    • applications to digit recognition
  • feature extraction and common feature types
    • isolated shape/object classification
  • Bayesian decision theory, maximum likelihood and Bayesian parameter estimation
    • presidential elections; disease testing; financial decision making
  • naive Bayesian classification
    • spam detection
  • linear discriminant methods, perceptron learning
    • computer security
  • k-means
    • image database visualization
  • hierarchical clustering
    • computational biology / phyolgeny
  • principal component analysis
    • image compression
  • multilayer perceptrons
    • object recognition
  • decision trees
    • medical decision making
  • pattern recognition with graphs
    • Google algorithm; social networks

These algorithms are widely used in many areas of computer science and will be illustrated on a variety of problems, including text and data mining, network security, image classification, OCR, and handwriting recognition.  The emphasis of the course is on a working understanding of algorithms and their properties, but also covers some important theoretical results. The course is accompanied by practical exercises in Numerical Python.

Document and Content Analysis

Tentatively being offered Summer 2009.

Hours: 2+1
Textbook: TBD
Course Language: English (lectures), English or German (exams, exercises)
Exercises: theory, Python, C++

Most of the data we interact with day-to-day does not come in the form of data structures or databases, but instead in the form of documents and document images.  This course introduces students to the formats, techniques, and algorithms used for representing, compressing, analyzing, processing, and displaying documents.  Topics covered include:

  • document formats and standards (TIFF, JPEG, PDF, PostScript, SVG)
  • document image compression (G4, MRC, token based compression, JPEG2000)
  • logical markup (HTML, XML, word processing formats, DocBook)
  • writings systems of the world
  • character sets and character encodings (ASCII, Unicode, special coding systems)
  • text rendering, layout, ligatures, and hyphenation (Pango)
  • typesetting and page layout systems (text flow, Word, LaTeX, etc.)
  • OCR (character recognition, page segmentation)
  • spelling and orthographic variation, statistical language modeling
  • document capture, page image dewarping and handheld document capture
  • named entity recognition, information extraction, table recognition
  • document search and retrieval, text mining, document databases
  • reading, psychophysics, and human-document interaction
  • document security and forensics

Privacy, Identity, and Computational Forensics

Tentatively being offered Winter 2009/2010

 Hours: 2+1 or 4+2
Textbook: TBD
Course Language: English (lectures), English or German (exams, exercises)
Exercises: Python, C

The course covers methods for computational analysis of digital data and physical evidence.  We will be examining how personal information and identity can be leaked and what techniques there are for protecting personal information.  Possible topics include:
  • interests groups: businesses, police, intelligence agencies, criminals
  • kinds of privacy and identity data
  • writer identification from text
  • handwriting identification from writing samples
  • printer identification from printouts
  • digital camera identification from images
  • social network analysis
  • pornography identification
  • forensic applications of image data
  • CAPTCHAs and reverse Turing tests
  • keystroke analysis for identity
  • biometric identification
  • cryptography, privacy, and identity
  • psychology and technology of phishing
  • multi-factor authentication
  • hard disk forensics
  • chain of custody and standards of evidence in forensics

Seminars & Labs

[LAB] Intelligent Data Mining and Pattern Recognition Competition

Usually offered every summer.

Every year, we pick a data mining, text mining, or pattern recognition competition and participate as part of a lab course.  The competitions are based on research challenges posed by companies, funding agencies, and public institutions.  Sample topics for these competitions include:
  • recognizing the content of news videos
  • detecting faces or objects in large image collections
  • identifying relationships in social networks
  • recommending music, movies, or products based on purchase histories
  • finding malignant cells in medical images
  • improving the quality of digital photographs
  • recognizing Arabic handwriting
  • identifying writers based on text or writing samples
  • hiding information in digital images--or detecting information hidden in digital images
The lab uses Python, Numerical Python, R, and C++, together with data mining, machine learning, and pattern recognition libraries.

[LAB] Computer Vision and Image Understanding

Usually offered every winter.

The lab course provides hands-on experience with computer vision and image understanding algorithms and forms the basis for research work in this area.    Topics vary slightly from year to year, but generally cover:
  • point matching and shape recognition using branch-and-bound algorithms
  • image database search using VQ codebooks and/or transform coefficients
  • discriminative learning
  • motion analysis and segmentation
  • stereo imaging and 3D surface modeling
  • multi-core programming
The lab is primarily based on Numerical Python, with OpenCV and other extensions; some C++ and Fortran 2003 programming.

Possible Future Courses

Bayesian Decision Theory and Statistical Learning

Tentatively being offered Summer 2010

Hours: 2+1 or 4+2
Textbook: TBD
Course Language: English (lectures), English or German (exams, exercises)
Exercises: theory, Numerical Python, R

Bayesian decision theory describes the fundamental principles of optimal decision making in the presence of uncertainty.  It has many practical applications in diverse fields such as finance, artificial intelligence, computer security, defense, and medicine. Bayesian decision theory also describes the basic laws that govern autonomous decision making in biological organisms, organizations, politics, evolution, games, and markets.

The course covers mathematical concepts and computational methods in Bayesian decision theory and statistical learning.  Topics tentatively include:

  • random variables, parametric models, Bayes rule
  • parametric models
  • Bayesian decision theory, risk, optimality, admissibility
  • priors and belief
  • hypothesis testing and frequentist methods
  • non-parametric and empirical Bayesian methods
  • Hidden Markov Models, Bayesian networks
  • variational Bayesian methods
  • Monte Carlo methods
  • causality
  • Bayesian machine learning
  • probably approximately correct learning, VC dimension
The course introduces students to the mathematical concepts and ideas underlying Bayesian decision theory, illustrates them with practical applications, and gives students computational tools to solve practical problems.

Simulations and Mathematical Modeling in Biology

TBD

 Hours: 2+1
Textbook: TBD
Course Language: English (lectures), English or German (exams, exercises)
Exercises: theory, Numerical Python, Fortran

The course will introduce students to a number of standard mathematical models and simulations (ODEs, PDEs, discrete event simulation, cellular automata models, finite element methods) in biological systems, including
  • diffusion
  • reaction/diffusion and pattern formation
  • predator/prey dynamics
  • ecological models
  • disease dynamics and epidemics
  • biological oscillators
  • birth-death processes
  • swarming
  • evolution and evolutionary computation
  • prisoner's dilemma and social cooperation
  • swarming behavior
  • genetic regulatory networks
  • protein dynamics
  • quorum sensing
  • bifurcation theory and chaos
The course aims to make abstract mathematical concepts accessible to biologists through simulation and visualization, while introducing mathematicians and computer scientists to important biological concepts.

[Seminar] Biologically Inspired Computing

Tentatively offered Winter 2009/2010.

Concepts from biology have formed the basis of many different approaches to computing and optimization.  What makes biological concepts attractive is that they tend to be intuitive and that biological systems are viewed as being robust and flexible.  In this seminar, we look at successful applications of biological concepts in computing, analyze their mathematical and theoretical foundations.  Possible topics include (depending on interest):
  • cellular computing and object-oriented programming
  • cellular automata and physical simulations
  • homeostasis and feedback mechanisms
  • neural networks for autonomous decision making
  • genetic algorithms and evolutionary computation for optimization problems
  • swarm algorithms and applications to computer networking
  • self-organization and applications to distributed sensor networks
  • immune system approaches to computer security
  • market models for automated and distributed  resource allocation
The seminar is aimed at computer scientists, and may also be of interest to mathematicians and biologists with interest and experience in computation.

Introductory Courses

The following courses are primarily for Bachelor students.

Intelligente Mensch-Maschine-Interaktion

  • Human-Centered Software Development
  • Visuelle GUI (Graphical User Interface) Designer und Prototyping Tools
  • Modelle und Prinzipien der menschlichen Wahrnehmung
  • Kulturelle Vielfalt und ihre Auswirkungen auf Benutzeroberflächen
  • Soziale Funktionen von Software
  • User Interface Testing and Evaluation
  • Sprache und Schrifterkennung
  • Benutzeroberflächen
  • Adaption
  • Menschliches Lernen
  • Web-Anwendungen
  • Information Retrieval und Knowledge Management

[Bachelor Seminar] Grundlagen der Mustererkennung [Bachelor]

Bachelor-Seminar: Einführung und Themenausgabe: Montag, 14. April in Raum 23-188 (Dr. Armin Stahl)

Voranmeldung wünschenswert per email an: courses@iupr.com

Ausgewählte Themen der Mustererkennung, z.B.

  • Klassifikation/Regression
  • Lineare Klassifikatoren
  • Nearest-Neighbor Klassifikation
  • Logistic Regression
  • Entscheidungsbäume
  • Supportvektormaschinen
  • Neuronale Netze