Automatic detection of colonic polyps and tumor in wireless capsule endoscopy images using hybrid patch extraction and supervised classification

Wireless Capsule Endoscopy (WCE) is an omnipotent noninvasive and painless diagnostic method for capturing digital images of entire Gastrointestinal (GI) tract. In this paper, we propose a method to detect colonic polyps and tumors from WCE images. Extractions of textural features are not only from single key point by utilizing single scale-invariant feature but also from neighborhood key points. Haralick texture features are extracted from each of patch size of 16*16 around the key points. For the best classification performance, the SIFT feature strategy is integrated with 22 Haralick textural features. In our prospective system, feature based classification is performed using Neural Network (NN) classifier for detecting colonic polyps and tumors accurately from the WCE images with an accuracy of about 97.5%.