Saarland University | Mathematical Image Analysis Group | Teaching
Winter term 2004 / 2005

Image Processing and Computer Vision

Prof. Joachim Weickert

Lectures (4h) with theoretical and programming assignments (2h); 9 ECTS points

Lectures: Tuesday and Friday, 11-13 c.t., Building 45, Lecture Hall 3
First lecture: Tuesday, October 26, 2004

New schedule starting Tuesday, Nov. 2:
Tuesday, 11-13 c.t., Building 45, Lecture Hall 2
Friday, 11-13 c.t., Building 45, Lecture Hall 3

Tutorials: 6 groups are scheduled for Monday 14-16, 16-18 and Friday 14-16, 16-18.

All tutorials take place in building 45.

Friday groups start Nov. 5, Monday groups start Nov. 8. Attendance is mandatory. Details see below.

The online registration for tutorials has been closed on Nov. 5. You can still use the group registration form to check which group you are currently registered for. For all questions concerning tutorial inscription contact Martin Welk, welk (at)

Type of Lectures

Broad introduction into mathematically well-founded areas of image processing and computer vision. These fields are important in numerous applications including medical imaging, computer-aided quality control, robotics, computer graphics, multimedia and artificial intelligence. The classes qualify for starting a bachelor thesis in our group.

The lectures are continued in the summer term by the in-depth course "Differential Equations in Image Processing and Computer Vision" which leads to current research topics. Both classes are necessary in order to pursue a diploma or master thesis in our group.


This course is suitable for students of mathematics or computer science. It counts either as a theoretical core course (Theorie-Stammvorlesung) in computer science or as an applied mathematics course. It is based on mathematical knowledge from the first two semesters. For the programming assignments, some elementary knowledge of C is required. The lectures are given in English.


From one week to the other, we alternate between programming and theoretical tutorials. The programming assignments give an intuition about the way that image processing and computer vision algorithms work, while the theoretical assigments provide additional mathematical insights. The tutorials are conducted by Andrés Bruhn (groups M1, M2, F1, F2), Martin Welk (group F4), and Stephan Didas (group F5).

Exercise sheets (programming and theoretical) are graded. Each exercise sheet allows to achieve a maximum of 12 points. Moreover, you are obliged to regularly attend the tutorials. Each time you miss a tutorial, 6 points are subtracted from your score (starting Nov. 5). To qualify for the exam, your final score (at the end of the term) must be 50 % or more of the possible points.

For additional preparation for the written exam, self-test problems have been prepared. Opportunity to ask questions related to these problems is given in the tutorials on Friday, Feb. 11, and Monday, Feb. 14, 2005.

Written Exam

The written exam took place on Tuesday (!), February 22, 2005 in the afternoon, from 3:00 to 6:00 p.m.

The following thresholds were applied in determining the grades:
Grade 1.0: minimum 57 points / 1.3: min. 54 p. / 1.7: min. 51 p. / 2.0: min. 48 p. / 2.3: min. 45 p. / 2.7: min. 42 p. / 3.0: min. 39 p. / 3.3: min. 36 p. / 3.7: min. 33 p. / 4.0: min. 29 p.
Note that the thresholds have been calculated based on a total of 59 points (instead of 64) taking into account the late announcement concerning problem 6c.

The detailed distribution of points and marks can be found here .

You could inspect your exam sheets on Thursday, March 17, between 2:30 and 4:00 p.m., building 27.2, basement, room 26 (opposite ladies' restrooms). Outside this time, individual inspection is possible on appointment in the offices of Andrés Bruhn, Martin Welk, Stephan Didas.

The second written exam took place on Thursday, March 31, 1:30 pm till 4:30 pm in the physics lecture hall (Großer Hörsaal Physik), building 22.
Everybody who was admitted to the first exam was automatically admitted to the second exam, independent on whether he/she had taken part, failed or passed the first exam. However, for those who have taken the second exam any grade from the first exam has become invalid – which implies that some grades have been improved, while some have been deteriorated, up to failing the whole course.

The point thresholds for the different grades were identical to those in the first exam.

The results can be queried via our online query form. Please note that the number of points and grades returned by the online query refer to the last exam you have written: Results from the first exam will be displayed if and only if you have not attended the second exam.

Inspection of exam sheets from the second exam: Friday, May 6, 1400–1500 h (2:00–3.00 p.m.), building 27.2, basement, room 26 (opposite ladies' restrooms).
Outside this time, individual inspection is possible only after previous e-mail appointment with Andrés Bruhn, Martin Welk, or Stephan Didas.


26. 10. Definitions, Image Types, Discretisation
29. 10. Degradations in Digital Images
2. 11. Image Transformations I: Continuous Fourier Transform
5. 11. Image Transformations II: Discrete Fourier Transform
9. 11. Image Transformations III: Image Pyramids
12. 11. Image Transformations IV: Wavelets
16. 11. Colour Perception and Colour Spaces
19. 11. Image Enhancement I: Point Operations
23. 11. Image Enhancement II: Linear Filtering
26. 11. Image Enhancement III: Wavelet Shrinkage, Median Filtering, M-Smoothers
30. 11. Image Enhancement IV: Mathematical Morphology
3. 12. Image Enhancement V: Diffusion Filtering
7. 12. Image Enhancement VI: Variational Methods
10. 12. Mathematical Aspects
14. 12. Image Enhancement VII: Fourier Methods und Deconvolution
17. 12. Feature Extraction I: Edges
21. 12. Feature Extraction II: Edges in Multichannel Images and Corners
11. 1. Feature Extraction III: Contour Representations and the Hough Transform
14. 1. Texture Analysis
18. 1. Segmentation I: Classical Methods
21. 1. Segmentation II: Variational Methods
25. 1. Image Sequence Analysis I: Local Methods
28. 1. Image Sequence Analysis II: Variational Methods
1. 2. 3-D Reconstruction I: Camera Geometry
4. 2. 3-D Reconstruction II: Stereo
8. 2. 3-D Reconstruction III: Shape-from-Shading
11. 2. Object Recognition I: Eigenspace Methods
15. 2. Object Recognition II: Moment Invariances
18. 2. Summary, Conclusions, Outlook

Material for the Programming Assignments

05. 11. Assignment 1: Fourier Transform
19. 11. Assignment 3: Point Operations
3. 12. Assignment 5: Morphology
17. 12. Assignment 7: Deblurring
21. 1. Assignment 9: Corners, Segmentation
4. 2. Assignment 11: Optic Flow

Self-Test Problems

Self-Test Problems (Problem Sheet)


There is no specific book for this class, but most topics are treated in one of the following books: These and further books can be found in the computer science library.

Furthermore, there is an interesting online compendium, where many researchers have written survey articles.
General informations and numerous links can be found at the Computer Vision Homepage. If you are looking for a specific reference, check out Keith Price's Annotated Computer Vision Bibliography. Numerous citations and online articles can be extracted from the CiteSeer webpage.

Joachim Weickert / April 28, 2005 / back to MIA home.
The author is not responsible for the content of external pages.
Imprint - Data protection