Keno Bressem — AI-Assisted Healthcare Lab

AboutÜber mich

Radiologist and researcher at the AI-Assisted Healthcare (AIAH) Lab, German Heart Center Munich, TUM. I work on AI methods for clinical medicine — mostly LLMs and computer vision for medical imaging.

Radiologe und Wissenschaftler an der AI-Assisted Healthcare (AIAH) Lab, Deutsches Herzzentrum München, TUM. Ich arbeite an KI-Methoden für die klinische Medizin — vor allem LLMs und Computer Vision für medizinische Bildgebung.

Medical Imaging NLP for Radiology Computer Vision Large Language Models Open Science

Selected PublicationsAusgewählte Publikationen

MRSegmentator: Robust Multi-Modality Segmentation of 40 Classes in MRI and CT Radiology: Artificial Intelligence, 2025
Evaluating the Effectiveness of Biomedical Fine-tuning for Large Language Models on Clinical Tasks Journal of the American Medical Informatics Association, 2025
Intermuscular Adipose Tissue and Lean Muscle Mass Assessed with MRI in People with Chronic Back Pain in Germany The Lancet Regional Health – Europe, 2025
Evaluating Reasoning Faithfulness in Medical Vision-Language Models using Multimodal Perturbations ML4H, 2025
Autonomous Medical Evaluation for Guideline Adherence of Large Language Models npj Digital Medicine, 2024
medBERT.de: A Comprehensive German BERT Model for the Medical Domain Expert Systems with Applications, 2024
Leveraging GPT-4 for Post Hoc Transformation of Free-text Radiology Reports into Structured Reporting Radiology, 2023

View all on Google Scholar → Alle Publikationen auf Google Scholar →

ProjectsProjekte

MRSegmentator

Robust segmentation of 40 anatomic structures on whole-body MRI with cross-modality generalization to CT.

Robuste Segmentierung von 40 anatomischen Strukturen in Ganzkörper-MRT mit modalitätsübergreifender Generalisierung auf CT.

LongHealth

QA benchmark with long clinical case documents to evaluate LLM reasoning over extended contexts.

Frage-Antwort-Benchmark mit langen klinischen Falldokumenten zur Bewertung des LLM-Reasoning über lange Kontexte.

gpt4-structured-reporting

Using GPT-4 to convert free-text radiology reports into structured formats for downstream analysis.

Nutzung von GPT-4 zur Umwandlung von Freitext-Radiologiebefunden in strukturierte Formate für weiterführende Analysen.

medAlpaca

Open-source medical LLM fine-tuned for clinical question answering. 550+ GitHub stars.

Open-Source-Sprachmodell für medizinische Fragen, feinabgestimmt auf klinische Daten. 550+ GitHub-Sterne.

ContactKontakt

keno.bressem@tum.de Google Scholar Google Scholar GitHub ORCID