Registration via Diffusion models

Deformable registration of medical images via diffusion denoising models 

Open Master Thesis

Contact person: Morteza Ghahremani [morteza.ghahremani@tum.de], Bailiang Jian, Christian Wachinger

 

There is an excellent opportunity in our team for master students to learn more about AI and medical imaging!
Here, at AI-Med (The laboratory for Artificial Intelligence in Medical Imaging; https://ai-med.de/), we are running an interesting project about deep learning-based deformable registration of the MRI image data. The task is mono-/multi-modal registration of MRI data, e.g. T1-weighted, T2-weighted, through state-of-the-art Diffusion Denoising Models (DDM). If you have a keen interest in AI & medical imaging and a strong desire to take an integrative approach to understanding the brain, we welcome you to join us!
Our offer to you:
  • We are offering a master’s thesis on “Deformable registration of medical images via diffusion denoising models”.
  • You will learn to know and apply the latest Deep Learning-based Deformable Registration methods.
  • You will learn to know and apply the sota Diffusion Denoising techniques in the field of medical image analysis.
  • You will be an active member of our international and interdisciplinary Research team.
  • You independently conduct research to find out how innovative ideas can be applied to add value.
What sets you apart:
  • You are studying computer science, medical technology, or a comparable course in a master’s degree.
  • You have gained experience with Python and PyTorch.
  • Experience with image registration or optical flow estimation would be a plus.
  • You have a good understanding of mathematics, deep learning and computer vision.
  • You would like to deal with demanding tasks in the field of medical technology.
Duration: 6 months
We look forward to receiving your application including a short statement of motivation, transcripts, and current CV.