(Pro-) Seminar Machine Learning and Computer Tomography

Thomas Schuster, Alice Oberacker
(Pro-) Seminar

News

Schedule

Written on 25.10.24 (last change on 06.11.24) by Alice Isabel Oberacker

Hello everyone,

All sessions will be held in E1.1 SR 4.12.

Below is the schedule for this semester.

  • Paper Assignment: Please verify that you have been assigned the correct paper and confirm your availability on your designated date.

  • … Read more

Hello everyone,

All sessions will be held in E1.1 SR 4.12.

Below is the schedule for this semester.

  • Paper Assignment: Please verify that you have been assigned the correct paper and confirm your availability on your designated date.

  • Initial Sessions: The first two sessions will follow a flipped classroom format, focusing on discussions and inquiries regarding Computed Tomography, Inverse Problems, and Machine Learning. 
    Kindly prepare by reading the script "Machine Learning for Inverse Problems" available in the Materials section. 
    Note that these sessions will not be traditional lectures; Prof. Schuster will be available to answer questions and facilitate discussions, rather than presenting the topics in a lecture format.

  • Presentations: Paper presentations will commence on November 27th and conclude on January 15th.

  • Submission Deadlines: Ensure that your slides/notes and handout are submitted to me by your assigned deadline.

  • Inquiries: Should you have any questions regarding the paper or your presentation, please feel free to contact me via email.

 

Schedule:

 
Date Student Paper title Deadline
6.11.24   Flipped Classroom  
13.11.24   Flipped Classroom  
27.11.24 Davide Damiano RecDNN: deep neural network for image reconstruction from limited view projection data 20.11.24
04.12.24 Ole Wolf A complementary l1-TV reconstruction algorithm for limited data CT 27.11.24
11.12.24 Mathivathana Ayyappan Solving ill-posed inverse problems using iterative deep neural networks 04.12.24
18.11.24 Philipp Henrikus Computationally Efficient Deep Neural Network for Computed Tomography Image Reconstruction  11.12.24
08.01.25 Kamna Dadwal LEARN: Learned Experts’ Assessment-based Reconstruction Network for Sparse-data CT 01.01.25
15.01.25 Norman Weinland Learned Primal-Dual Reconstruction 08.01.25

 

Available papers

Written on 17.10.24 (last change on 21.10.24) by Alice Isabel Oberacker

Hi,

I will update you here on which papers are still available.

The following papers can still be selected:

  • An Educated Warm Start For Deep Image Prior-Based Micro CT Reconstruction
  • Computationally Efficient Cascaded Training for Deep Unrolled Network in CT… Read more

Hi,

I will update you here on which papers are still available.

The following papers can still be selected:

  • An Educated Warm Start For Deep Image Prior-Based Micro CT Reconstruction
  • Computationally Efficient Cascaded Training for Deep Unrolled Network in CT Imaging
  • Superiorization-inspired unrolled SART algorithm with U-Net  generated perturbations for sparse-view and limited-angle CT reconstruction
  • Structure-aware diffusion for low-dose CT imaging
  • Deep Convolutional Neural Network for Inverse Problems in Imaging
  • Conditional Invertible Neural Networks for Medical Imaging
  • Computed tomography reconstruction using deep image prior and learned reconstruction methods
  • One Sample Diffusion Model in Projection Domain for Low-Dose CT Imaging
  • Super-Resolution and Sparse View CT Reconstruction
     

Paper selection

Written on 16.10.24 by Alice Isabel Oberacker

Hello everyone,

The papers are now uploaded to the materials section in CMS.

Once you decided which paper interests you, please send me an E-Mail with the title and any preferred dates.

 

All the best,

Alice

 

Machine Learning and CT

  1. Choose Your Paper: Select your preferred paper by October 23rd. Send an email to Alice to let her know your choice. You'll find all the papers and the script for "Machine Learning for Inverse Problems" in the materials section.

  2. Presentation Requirements:

    • Your presentation must be 45 minutes for a pro-seminar and 60 minutes for a seminar.

    • After each presentation, we'll have a discussion and a Q&A session.

    • Prepare slides for your presentation. If you prefer a blackboard-only presentation, ensure you have detailed notes ready.

    • Create a handout summarizing your topic, with a maximum length of 3 pages.

    • Submit both your slides and handout to Alice at least one week before your assigned presentation date.

    • Regular attendance is required to pass this course successfully.

  3. Session Details:

    • The first session will take place on November 6th.

    • Prof. Schuster will hold a flipped classroom session discussing Inverse Problems and Machine Learning.

    • Every following session will be a student presentation.

 

 

 

Hello everyone,

in the upcoming winter semester, we are offering a seminar/proseminar on "Machine Learning and Computed Tomography" (7/5 CPs respectively).

We offer presentations on current topics such as ML methods for low dose and sparse view CT. The seminar is aimed at Bachelor and Master students of mathematics, mathematics and computer science, financial mathematics as well as teacher training students. The first meeting will take place in the first week of the winter semester (starting October 14).

We hope that many of you will be interested.

Best regards,

Thomas Schuster

 

 

Dozent: Prof. Dr. Thomas Schuster

Vorlesungszeiten: Mittwoch  10:15-11.45 Uhr
Vorlesungsort: E1.1 SR 4.12

 

Sprechstunden / Office hours

Alice Oberacker: auf Anfrage - on request

Thomas Schuster: auf Anfrage - on request

 

Fragen / Questions

Bei Fragen wenden Sie sich bitte an Alice Oberacker.

If you have any questions, please contact Alice Oberacker.

 

Privacy Policy | Legal Notice
If you encounter technical problems, please contact the administrators.