Guest Lecture Series Medical Information Sciences

Veranstaltungsort: Hörsaal N2045 (Fakultät für Angewandte Informatik)

Abstract

Artificial intelligence (AI) is increasingly transforming gastrointestinal diagnostics and therapy. I will provide a comprehensive overview of state-of-the-art AI-based methods in endoscopy, with a particular focus on the early detection and treatment of Barrett’s-associated carcinoma.

Using clinically relevant examples, I will demonstrate how deep learning models are applied to detection, classification, and semantic segmentation of endoscopic image data. A key emphasis is placed on semi-supervised learning approaches, which enable the training of robust models despite the limited availability of annotated medical data. In this context, error-correcting supervision and Mean Teacher–based frameworks are presented as effective strategies to improve model consistency, robustness, and clinical reliability.

In addition, I will address the role of the human-in-the-loop, challenges related to trust in AI systems, and the problem of overconfident predictions, including methods for out-of-distribution detection in medical imaging. Recent clinical studies show that AI can significantly improve diagnostic accuracy when used in combination with expert judgment or as a standalone system.

Finally, AI-supported assistance systems for endoscopic submucosal dissection (ESD) are presented, including real-time overlays, phase recognition, and safety alerts, highlighting the successful translation of AI research into clinical practice. The lecture thus bridges methodological foundations, clinical evidence, and practical deployment of intelligent systems in gastrointestinal medicine.

Referent:   Prof. Dr. Christoph Palm

Kurzbiographie

Christoph Palm is Professor of Medical Image Computing and Artificial Intelligence at OTH Regensburg, Germany, where he heads the Regensburg Medical Image Computing (ReMIC) lab. He received his Diploma degree in Computer Science and his Dr. rer. nat. (PhD) from RWTH Aachen University. Prior to his appointment in Regensburg, he held research positions at Research Center Jülich, University College London, and RWTH Aachen, and worked in industry as Head of Machine Translation at Aixplain AG.

His research focuses on medical image computing, machine learning, and deep learning, with a particular emphasis on computer-aided diagnosis, smart endoscopy, and image-based assistance systems for clinical applications. Prof. Palm has published more than 120 peer-reviewed scientific papers in leading conferences and journals such as CVPR, ECCV, MICCAI, Computers in Biology and Medicine, GUT, and Endoscopy. His work has received several best paper and challenge awards. He is actively involved in the international scientific community as a conference organizer, program committee member, and reviewer for high-impact journals and conferences. Since 2024, he holds the right to award doctorates and is a member of the Board of Directors of the Regensburg Center for Artificial Intelligence at OTH Regensburg.

Further information will be forwarded with our next circular email and can already be found on our guest lecture series’ official home page.