Cybernetics Seminar – “From Compressed Sensing to deep learning-based MR image reconstruction“

Dr. Thomas Küstner – Friday, February 13th, at 11:00 (CEST)

This Friday, February 13 at 11am, the Cybernectic Seminar will have an in-person talk by Dr. Thomas Küstner at the Max Planck House Lecture Hall (Max-Planck-Ring 6, 72076) and via Zoom!

Zoom link to attend the seminar:

https://eu02web.zoom-x.de/j/8525355685?pwd=bzJKWkE3ckFiSTcyV0hxc0NVcGJGQT09

Meeting ID: 852 535 5685

Password: 468631

Dr. Thomas Küstner Talk Information:

Title: From Compressed Sensing to deep learning-based MR image reconstruction

Abstract:

Magnetic resonance imaging (MRI) reconstruction is a fundamental inverse problem, where undersampling of k-space is applied to reduce scan time while preserving diagnostic image quality. Traditionally, this challenge has been addressed using model-based reconstruction techniques such as parallel imaging, compressed sensing, and low-rank modeling, which incorporate explicit prior assumptions through handcrafted regularizers. Over the past decade, deep learning has profoundly reshaped this field by enabling data-driven regularizations. This presentation provides a structured overview of the evolution from classical compressed sensing frameworks to modern deep learning–based MRI reconstruction. We review representative clinical applications and critically address remaining challenges such as generalizability, domain shifts, hallucinations, and reliability.

Bio:

Dr. Thomas Küstner received his PhD from the University of Stuttgart in 2017. From 2018 to 2020, he held a postdoctoral fellowship at the School of Biomedical Engineering and Imaging Sciences at King’s College London. Since 2021, he has led the Medical Imaging and Data Analysis group (MIDAS.lab) at the University Hospital of Tübingen. His research group focuses on AI applications for multiparametric and multimodal imaging techniques (MRI, CT, PET) in acquisition and reconstruction, as well as the analysis of clinical and epidemiological studies. His focus is on AI developments in MRI, particularly in the areas of accelerated acquisition, motion correction, reconstruction, and image quality analysis.

The cybernetics seminar provides us a fantastic opportunity to speak with visiting researchers. Please contact gabriele.lohmann@tuebingen.mpg.de if you’re interested in meeting Dr. Thomas Küstner 1-on-1!

You can find information on the talk below. See you there!