Welcome to the homepage of the lecture

Convex Analysis for Image Processing

Winter Term 2013/2014

Convex Analysis for Image Processing

Lecturer: Laurent Hoeltgen,
Office hours: Tuesday, 14:00-16:00
Examiner: Prof. Dr. Joachim Weickert
Lectures (3h) with exercises (1h)
(6 credit points)

Time and Location:
Monday 14-16 in Building E1 3, HS 003
Thursday 12-14 in Building E1 3, HS 003

First lecture: Thursday, October 17, 2013

AnnouncementsDescriptionPrerequisitesLecture notesAssignmentsExamsLiterature

  • The registration for the lecture is now possible. For further details, see below.
  • Important details concerning the registration for the exam can be found below.
  • The Lecture on November 25th will be held in HS 002 instead of HS 003.
  • The list of students admitted to the exam is online.

Many problems in image processing can be modeled as convex minimization problems. The restored image is thus the minimizer of a suitable convex energy functional. We will cover the foundations of convex analysis (convex sets and functions, subdifferentials, duality, optimality conditions) and their applications to nonsmooth nonlinear optimization.

Except for basic concepts of calculus and linear algebra this course is self-contained, especially, because we will be working in the finite-dimensional setting. There will also be programming assignments which illustrate the relevance of fast optimization algorithms.

Undergraduate knowledge in mathematics (e.g. ''Mathematik für Informatiker I-III''). The lectures will be given in English.

The script can be downloaded here. It will be updated continuously during the semester.

DateLecture/Tutorial
2013-10-14No lecture
2013-10-17Lecture
2013-10-21Lecture
2013-10-24Tutorial
2013-10-28Lecture
2013-10-31Lecture
2013-11-04Lecture
2013-11-07Tutorial
2013-11-11Lecture
2013-11-14Lecture
2013-11-18Lecture
2013-11-21Tutorial
2013-11-25Lecture
2013-11-28Lecture
2013-12-02Lecture
2013-12-05Tutorial
2013-12-09Lecture
2013-12-12Lecture
2013-12-16Lecture
2013-12-19Tutorial
2014-01-06Lecture
2014-01-09Lecture
2014-01-13Lecture
2014-01-16Tutorial
2014-01-20Lecture
2014-01-23Lecture
2014-01-27Lecture
2014-01-30Lecture
2014-02-03Lecture
2014-02-06Lecture

Homework will be assigned bi-weekly. To qualify for the exam you need 50% of the points from these assignments. Starting with the second assignment sheet, submissions in groups of at most 2 people is possible. The list with the Student IDs admitted to the exam can be found here. If you think, this list is wrong, pleas contact Laurent Hoeltgen as soon as possible.

Due dateAssignmentAdditional Material
2013-10-24Assignment 1-
2013-11-07Assignment 2-
2013-11-21Assignment 3-
2013-12-05Assignment 4-
2013-12-19Assignment 5Plot for Assignment 3
2014-01-16Assignment 6Material for the programming assignment, Example Solution
2013-12-05Optional Programming Assignment 1Example Solution in Fortran

The registration for the exam in the HISPOS system is possible from October 31st, 2013 until November 17th, 2013. In order to participate in the exam, it is mandatory to be enlisted to the exam in the HIPOS system. If you have difficulties registering, do not hesitate to contact studium@cs.uni-saarland.de.

The first exam will be on the 7th February 2014 from 10:00 to 12:00 in HS003 Bldg. E1 3. There will also be a second exam on the 26th March 2014 from 10:00 to 12:00 in HS003 Bldg. E1 3. These will be closed books exams lasting 2 hours.

In order to register for the Lecture, write an e-mail to Laurent Hoeltgen. The subject line must begin with the tag [CAIP13]. Further, please provide the following information: first name, last name, date of birth, student ID number (e.g. Matrikel), course of study (e.g. Bachelor, Master, ...), subject of your studies (e.g. computer science, mathematics, ...). These information are necessary to give you credit points for a successful participation in the lecture at the end of the semester. Please ensure that the provided information is correct. Note that the e-mail address from which you send this information will be used to provide you with urgent information concerning the lecture. Therefore, use an address, that you check regularly. Finally, this registration is for internal purposes at our chair only and completely independent of any System like LSF/Hispos. They require a separate registration.

Please remember that you have to register online for the lecture/exam in the HISPOS system of the Saarland University

The lecture notes which will be provided for this course are self-contained. No textbook is required. Examples of books giving background material and further reading are:

  1. Convex analysis
    R. T. Rockafellar, Princeton University Press, 1997
    (Classical text on convex analysis.)
  2. Convex analysis and minimization algorithms I+II
    J.-B. Hiriart-Urruty and C. Lemaréchal, Springer, 1993
    (Provides many nice examples and motivational ideas.)
  3. Convex Optimization
    S. Boyd and L. Vandenberghe, Cambridge University Press, 2004
    (Also available to download at the authors' webpages.)
  4. Perturbation analysis of optimization problems
    Joseph F. Bonnans and A. Shapiro, Springer, 2000
    (Works in the infinite-dimensional setting. Nice introduction to duality.)
  5. Numerical Optimization
    J. Nocedal and S. J. Wright, Springer, 2006
    (Introduction to a wide range of modern optimization techniques. Good motivational examples.)
  6. Image Processing and Analysis
    T. F. Chan and J. Chen, SIAM, 2005
    (Not only restricted to optimization methods in image processing but introduces major models.)


Last change: Laurent Hoeltgen 17.01.2014

MIA Group
©2001-2020
The author is not
responsible for
the content of
external pages.

Imprint - Data protection