Deconvolution Methods in Image Processing

Summer Term 2006

Deconvolution Methods in Image Processing

Dr. Martin Welk, Prof. Dr. Joachim Weickert


Summer 2006

Seminar (2h)

Important DatesDescriptionAdministrative detailsOverview of Topics

First meeting: Monday, February 13, 2006, 16:15h (4:15 p.m.), Bld. E11, room 306

Sign-up deadline: Friday, February 10, 2006 (elapsed). No free capacity is left in this seminar.

Regular meetings during summer term 2006: Wednesdays, 16:15h (4:15 p.m.), Bld. E11, room 306
No meeting on June 7!

Please direct all organisatorial requests to Martin Welk.

Contents: In the acquisition of images, blurring is a ubiquitous phenomenon. An important type of blurring is mathematically described by convolution of a sharp image with a point-spread function.

The sharpening of so degraded images, deconvolution, is an ill-posed and thus difficult problem in image processing. A large variety of deconvolution methods have been developed during the last decades, which differ in their assumptions and which offer different advantages and disadvantages.

In this seminar, an overview over a broad range of deconvolution methods will be given, stretching from direct inversion approaches via iterative solution of large systems of equations to variational models combining deconvolution with other tasks like segmentation or image completion (inpainting).

Prerequisites: The seminar is designed for graduate students (i.e. advanced Bachelor, Master, and post-Vordiplom) of computer science or mathematics. For most topics, completed MfI courses are sufficient.

Language: Understanding scientific text in English is indispensable. Talks and write-up can be in English or German.

Sign-up: closed. The deadline was February 10, 2006. No free capacity is left.

Sessions: The initial meeting took place on Monday, February 13, 2006, 16:15h (4:15 p.m.), Bld. E11, room 306.
Regular sessions will be during the summer term on Wednesdays, 16:15h (4:15 p.m.), Bld. E11, room 306. The date of the first meeting will be communicated by e-mail. During that meeting, the further schedule will be determined.
Participants who are hindered to attend a meeting for substantial reasons should notify M. Welk in advance by e-mail.

Requirements for successful participation («Schein»): regular attendance, talk, write-up

Please note that the presentation dates given here are preliminary and subject to change.

No. Topic
Presented by
1 Iterative methods for image deblurring
Michael Maurer, Daniel Dahrendorf, Peter Geber[PPT]
April 19
2 Deblurring Gaussian blur
Christiane Altmeyer
April 26
3 Analytic inverse of discrete Gaussian blur
Adrian Bock[PPT]
April 26
5 Blur identification by generalised cross-validation
Pascal Gwosdek
May 3
No session on May 10
4 Blur identification by residual spectral matching
Süntje Böttcher
May 17
6 Blind deconvolution by means of the Richardson-Lucy algorithm
Kim Pecina
May 17
7 Matrix-algebraic iterative regularisation methods
Kristina Scherbaum, Harriet Bach
May 24
8 Regularised blind deconvolution using recursive inverse filtering
Shu Peng
May 31
9 Deconvolution with correct sampling
Yana Mileva[PPT]
May 31
No session on June 7
10 Total variation based deblurring
Markus Mainberger, Martin Grochulla
June 14
11 A regularisation approach to joint blur identification and image restoration
Thomas Bühler
June 21
12 Blind deconvolution of bar-code signals
Jean-Marc Roth
June 21
13 PDE-based deconvolution with forward-backward diffusivities and diffusion tensors
Mohamed Abdelmaksoud
June 21
14 Variational pairing of image segmentation and blind restoration
Han Dong
June 28
15 Image deblurring in the presence of salt-and-pepper noise
Gerrit Kahl
June 28
17 Combining TV and wavelet packet approaches for image deblurring
Marin Katov
July 5

Martin Welk / April 5–July 31, 2006

MIA Group
The author is not
responsible for
the content of
external pages.

Imprint - Data protection