Medical Imaging

Medical imaging plays a central role in disease detection, diagnosis, and treatment planning. However, image quality often comes at a cost: higher doses of radiation. To reduce patient exposure, especially in modalities like X-ray, mammography, digital breast tomosynthesis (DBT), and cone-beam computed tomography (CBCT), clinicians must rely on low-dose imaging protocols. While these protocols minimize risk, they introduce significant challenges—most notably increased image noise, blur, and loss of anatomical detail.

Our research group is dedicated to advancing deep learning solutions that intelligently enhance the quality of medical images while maintaining the benefits of low-dose acquisition. By applying cutting-edge AI models, we aim to:

• Denoise images acquired with minimal radiation

• Suppress artifacts from undersampled or distorted acquisitions

• Preserve fine anatomical features critical for accurate diagnosis

• Improve reliability and confidence in automated decision-making

We apply these technologies across a wide range of areas—from breast imaging and 3D scans to disease classification (like COVID-19) and detailed anatomical mapping. Our goal is not just better images, but smarter tools that help healthcare professionals make confident decisions, even in challenging environments.

By combining deep learning with clinical insight, we aim to build a new generation of imaging technology that is safer for patients and more powerful for providers. We're always looking for collaborators who share our vision of advancing healthcare through intelligent, responsible innovation.

 

Collaborations

• Siemens Healthineers
• STIMULATE
• University Hospital Magdeburg

 

Selected Publications

• An Interpretable X-Ray Style Transfer Via Trainable Local Laplacian Filter.
  Dominik Eckert, Ludwig Ritschl, Christopher Syben, Christian Hümmer, Julia Wicklein, Marcel Beister, Steffen Kappler & Sebastian Stober.
  In: IEEE 22nd International Symposium on Biomedical Imaging (ISBI). IEEE, 2025.
  [Paper]
• Deep learning based tomosynthesis denoising: a bias investigation across different breast types.
  Dominik Eckert, Julia Wicklein, Magdalena Herbst, Stephan Dwars, Ludwig Ritschl, Steffen Kappler & Sebastian Stober.
  In: Journal of Medical Imaging, 2025.
  [Paper]
• Towards Patient Specific Reconstruction Using Perception-Aware CNN and Planning CT as Prior.
  Suhita Ghosh, Philipp Ernst, Georg Rose, Andreas Nürnberger & Sebastian Stober.
  In: IEEE 19th International Symposium on Biomedical Imaging (ISBI). IEEE, 2022.
  [Paper]
• Exploration of interpretability techniques for deep COVID-19 classification using chest X-ray images.
  Soumick Chatterjee, Fatima Saad, Chompunuch Sarasaen, Suhita Ghosh, Valerie Krug, Rupali Khatun, Rahul Mishra, Nirja Desai, Petia Radeva, Georg Rose, Sebastian Stober, Oliver Speck & Andreas Nürnberger.
  In: Journal of Imaging (AI in Imaging), 2024.
  [Paper]
• Uncertainty-aware temporal self-learning (UATS) - semi-supervised learning for segmentation of prostate zones and beyond.
  Anneke Meyer, Suhita Ghosh, Daniel Schindele, Martin Schostak, Sebastian Stober, Christian Hansen & Marko Rak.
  In: Artificial intelligence in medicine: AIM - Amsterdam [u.a.]: Elsevier Science - AIM, Bd. 116, 2021.
  [Paper] [Code]