Computational Methods for Imaging

Winter Semester 2019/20

This seminar will cover different topics for mathematical imaging depending on the prior knowlegde of the participants. Possible topics are for example different inverse problems in imaging applications and variational regularization methods leading to optimization problems.

The semiar will be a block seminar. The concrete dates of the seminar will be discussed together with the participants in a first meeting in October.

Time schedule and topics

• 12 December 2019, room 415
  • Linear filters and convolution
  • Windowed Fourier transform and continuous wavelet transform – cancelled!
  • Discrete wavelet transform and image compression
  • Bayesian inversion framework
  • Optical flow estimation
• 10 January 2020, lecture hall H3
  • Nonlinear diffusion – Perona Malik equation – postponed to 17 January 2020
  • Denoising images – wavelet shrinkage denoising
  • Denoising images – Total variation denoising – postponed to 17 January 2020
• 16 January 2020, room 415
  • Sparsity-promoting regularization
  • TV regularized reconstruction via Chambolle-Pock algorithm
  • Nonlinear inverse problems
  • Image inpainting – cancelled!
• 17 January 2020 (alternative date), room 127
  • Nonlinear diffusion – Perona Malik equation
  • Denoising images – Total variation denoising

Literature:

• Bredies and Lorenz, Mathematical Image Processing
• Aubert and Kornprobst, Mathematical Problems in Image Processing
• Engl, Hanke, Neubauer, Regularization of Inverse Problems