A data mining algorithmic approach for processing wireless capsule endoscopy data sets

Wireless capsule endoscopy (WCE) has been a breakthrough in recent medical technology. It is used to view the gastrointestinal tract and detect abnormalities such as bleeding, Crohn's disease, peptic ulcers, and colon cancer. In this paper data mining techniques are utilized to extract useful information from a dataset of abnormal regions and non-abnormal regions. More specifically, the dataset contains polyps regions, ulcers regions and healthy regions. A number of features (shape descriptors, texture descriptors and color information) has been extracted for these regions and using a data mining toolbox useful conclusions are given on various relationships between these regions.