Interactive Segmentation of Capsule Endoscopy Images Using Grow Cut Method

Abstract

High incidence of Gastrointestinal (GI) tract related diseases are common nowadays, which leads to cancer. Diagnose of these kind of diseases is not easy in the early stage. This makes the researchers to develop an automated screening technique for early detection, many modalities are employed to view the GI tract but complete visualization is not possible. Wireless Capsule Endo scopy (WCE) is a modern modality which helps to view the complete GI tract, This claims an computational assistances which automatically segment the defective frames with the help of computer assisted diagnosis, In this paper the author describes an semi-automated interactive grow cut algorithm for segmenting capsule endoscopy (CE) images as foreground and background image by iterative process with a small number of user-labeled pixel the rest of the images are segmented with less human input. User can view and guide the algorithm without any additional user effort a semi automated segmentation is computed using mat lab.