Fraunhofer MEVIS is participating with numerous scientific presentations at the 29th International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI) from September 27 to October 1, 2026, in Strasbourg, France. The annual MICCAI conference serves as a forum for the world’s leading biomedical researchers, engineers, and clinicians from a broad spectrum of disciplines in the fields of medical imaging and computer-assisted intervention.

Researchers of Fraunhofer MEVIS are contributing to the following sessions, challenges, and workshops:

Doctoral Symposium

• Kai Geissler presents a poster on “Deep Learning for Automated Reporting in Breast Magnetic Resonance Imaging”, Sep 28, 16:00–18:00

Challenges

• HECKTOR – Head and Neck Tumor Lesion Segmentation, Staging, and Prognosis: Kai Geissler has participated in this challenge with “Unified vs modular methods for head-neck tumor and lymph node segmentation, staging, and survival prediction”, Sep 27, 08:00–18:00
• MVAA – Mitral Valve Anatomy Analysis Using Multimodal Imaging Data: Julia Mellert (Charité), Lisa Bautz, and Johanna Brosig have participated in this challenge and present a poster on “Multimodal Mitral Valve Segmentation: Mesh-Based 3D and Depth-Transfer Learning”, Sep 27, 13:30–15:30

Workshops

• FAIMI – Fairness of AI in Medical Imaging: Co-organized by Eike Petersen, Sep 27, 10:30–18:00
• iMIMIC – Interpretability of Machine Intelligence in Medical Image Computing: Poster by Peachapong Poolpol et al. on "Faithful Faithfulness Evaluations: Challenges & Pitfalls Learned from a Breast MRI Case Study", Sep 27, 15:30–16:00
• MLMI – Machine Learning in Medical Imaging: Poster by Mohamed Amine Kina et al. on "Prevalence calibration as shortcut mitigation", Sep 27, 15:30–16:00
• EMA4MICCAI – Efficient Medical AI: Poster by Kai Geissler on “RadYOLO: Computationally efficient 3D Object Detection and Segmentation in CT and MRI” and Poster by Felix Thielke on "Segmentation Without Full Logits: Fused Kernels for Memory-Efficient Many-Class Models" , Oct 1, 10:00–10:30

RadYOLO is a highly efficient detection model for three-dimensional radiological images developed at Fraunhofer MEVIS. It is available as open-source software at: https://github.com/FraunhoferMEVIS/RadYOLO