Real-Time Mapping from Endoscopic Video

Mapping of regions of importance is fundamental to interact with the environment, from satellite maps of glaciers to functional maps for GPS navigators. Autonomous vehicles also work by matching visual data to an existing map which needs to be constantly updated and extended. Now, EndoMapper is developing a similar mapping system to support endoscopic procedures like colonoscopy, tumour biopsy, or even targeted drug delivery. To accomplish their world-first, the team must tackle the challenge of modelling non-rigidity. Unlike roads and byways, the body's tissues and organs are prone to change as the endoscope moves. Taking advantage of new non-rigid mathematical models and machine learning, endoscopy may soon enter the realm of augmented virtual reality and autonomous navigation.

Logo European Comission