Novel Convoluted Local Energy Oriented Patten(CLEOP) for the Classification of Wireless Capsule Endoscopy images

Endoscopy enables Physician to identify inflammation, ulcers, and tumors by viewing the intestines and other organs of the digestive systems of a human being. Upper endoscopy examination inside digestive system gives better accuracy when compared to X-rays in detecting abnormal growths caused by cancer and other diseases. Automation in medical image processing has enhanced the prediction of affected layer from the source of image. In that, image analysis in endoscopic medical field was focused on high level prediction and to estimate proper and better treatment for patients at earlier stage with less amount of stress. There are several methods of image analysis to extract the features of the image and predict its category. For better prediction model, the features of that image should be rotational invariant to find the best match training set. To improve the feature extraction in image processing application, texture patterns performed with better efficiency to analyze an image. In this proposed work, a novel method of texture pattern analysis method employed to classify the endoscopic image. This was achieved by using the Convoluted Local Energy Oriented Pattern (CLEOP) based feature extraction method. For validating the performance of proposed work, the CVC-ClinicDB image database was used in the testing process. In that, the database was separated into two major types based on the level of affected tissue region and its size. The efficiency of CLEOP texture pattern extraction method is compared with other state-of-the-art methods of feature extraction and classification model. This can be justified by estimating the statistical parameters estimated by validating the ground-truth of image database.