SiMucosa: AI-based endoscopy simulation on silicone models

Colonoscopy simulators are a cornerstone of endoscopic training, but their artificial silicone appearance has long been recognised as a limitation. What trainees see on screen looks little like a real colon, which reduces how well the experience prepares them for actual procedures.

SiMucosa addresses this directly. Powered by EndoStyle, our AI-driven style transfer system, it converts the video feed of a standard silicone simulator into realistic-looking mucosal imagery in real time, with no changes to the physical device or how it is handled. 

The result is a familiar training tool that now looks the part, bringing the on-screen experience meaningfully closer to clinical reality.

 

We thank Dr Henke Medical Training for providing us with silicone-based endoscopy phantoms.

Funding

This research project is funded by the Eva Mayr-Stihl Foundation.

Logo Eva Mayr-Stihl Stiftung

Contact

Portraitfoto von Prof. Dr. med. Alexander Hann

Prof. Dr. med.
Alexander Hann

Head of the working group

+49 931 201-45918

Portraitfoto von Joel Troya


Joel Troya

Hardware Engineer

Portraitfoto von Jana Theile


Jana Sophia Theile

Physician and Research Associate

Portraitfoto von Irshath Ali Syed


Irshath Ali Syed

Hardware Engineer

Address

Medizinische Klinik und Poliklinik II | Gastroenterologie | Oberdürrbacher Straße 6 | Haus A4 | 97080 Würzburg | Deutschland

Auvera Haus | Grombühlstraße 12 | 97080 Würzburg | Deutschland