Impact of image segmentation techniques on celiac disease classification using scale invariant texture descriptors for standard flexible endoscopic systems

Abstract

Celiac disease (CD) is quite common and is a proximal small bowel disease that develops as a permanent intolerance to gluten and other cereal proteins in cereals. It is considered as one of the most difficult diseases to diagnose. Histopathological evidence of small bowel biopsies taken during endoscopy remains the gold standard for diagnosis. Therefore, computer-aided detection (CAD) systems in endoscopy are a newly emerging technology to enhance the diagnostic accuracy of the disease and to save time and manpower. For this reason, a hybrid machine learning methods have been applied for the CAD of celiac disease. Firstly, a context-based optimal multilevel thresholding technique was employed to segment the images. Afterward, images were decomposed into subbands with discrete wavelet transform (DWT), and the distinctive features were extracted with scale invariant texture recognition. Classification accuracy, sensitivity and specificity ratio are 94.79%, 94.29%, and 95.08%, respectively. The results of the proposed models are compared with the results of other state-of-the-art methods such as a convolutional neural network (CNN) and higher order spectral (HOS) analysis. The results showed that the proposed hybrid approaches are accurate, fast, and robust.