Automatic Discrimination of Duodenum in Wireless Capsule Video Endoscopy
ABSTRACT Wireless Capsule Video Endoscopy is a recent acquisition method providing an internal view of the gastrointestinal tract which
is currently applied in a large quantity of methods for detecting different intestinal diseases. In some of these applications,
the automatic identification of some regions of the small intestine is essential. However, the high amount of time needed
for video visualization makes this task unfeasible. In this paper, we present a novel system for automatical labelling the
transition from proximal to distal parts of the small bowel in the capsule endoscopy video based on textural descriptors.
Results show an accuracy of the proximal-distal boundary detection of more than 70%.