Identification of polyps in Wireless Capsule Endoscopy videos using Log Gabor filters

Abstract

Wireless Capsule Endoscopy (WCE) is a very useful technology that enables gastroenterologists to examine the human digestive tract and more particularly, the small bowel, searching for various abnormalities like blood-based abnormalities, ulcers and polyps. Each WCE video consists of approximately 50,000 frames making its examination a very tedious task. In this paper a methodology is proposed for the automatic detection of polyps. It utilizes the very promising Log Gabor filters as a segmentation scheme, in conjunction with the powerful SUSAN edge detector. After segmentation process, geometric information of the resulted segments is extracted to identify polyp candidates. Furthermore, certain rules apply to limit the number of false positive detections. Illustrative and statistical results of the methodology are also given in this paper.