Seminar: Computational Methods for Image Reconstruction

Administrative information

Seminar course for Bachelor students (IN0014) and Master students (IN2107).

• Organizer: Tobias Lasser
• Sessions:
  • kick-off session on Oct. 18, 2023 (10:15 to 11:45), room 1.211 in MIBE building (Boltzmannstr. 11, Garching)
  • two block sessions on Jan. 10 and 11, 2024 (9:00 to 17:00), room 1.211 in MIBE building (Boltzmannstr. 11, Garching)
• Course language: English

Registration

Registration is closed.

Course overview

Image restoration, image denoising, X-ray Computer Tomography (CT), or Single Photon Emission Computed Tomography (SPECT) are example applications of image reconstruction methods. These image reconstruction methods see widespread use in industrial and commercial settings, as well as in the medical domain to aid diagnosis and therapy planning. The underlying computational methods are the focus of this seminar course, ranging from discretization, formulation as an optimization problem for particular application domains, regularization, optimization algorithms, to image quality assessment. In this seminar course, we will study those methods in sufficient detail to get an insight into the mathematics and algorithms employed in image reconstruction. This course is based on the book Image Reconstruction: Algorithms and Analysis by Jeffrey A. Fessler.

Course modalities

The seminar course will take place in two full-day block sessions in presence on January 10 and 11, 2024. A kick-off meeting at the beginning of the term (October 18, 2023) will serve to bring everyone up to speed on the seminar's requirements.

Each participating student will focus on one of the chapters of the book the course is based on (see above). The students will present the contents of the chapter in a 60 minute lesson. This lesson will be split into an interactive presentation of the chapter contents (45 minutes), and a presentation of the code that (s)he has written to support that chapter's contents while showing corresponding image results (15 minutes).

One week after the block sessions, a written report in PDF format has about the lesson has to be submitted, along with a repository containing the associated code.

Aims of the course

After participating in this seminar course, students will be able to understand the basics of computational methods used for image reconsctruction. In addition, the students will have implemented key concepts of those computational methods on their own, gathering some hands-on experience in image reconstruction methods, for example for X-ray Computed Tomography or Image restoration.

Prerequisites

Mathematical knowledge from Bachelor studies is required, along with some programming experience in a language like Python or Matlab.