A multi-scale comparison of texture descriptors extracted from the Wavelet and Curvelet domains for small bowel tumor detection in Capsule Endoscopy exams

摘要

Traditional endoscopic methods do not reach the entire GI tract. Wireless Capsule Endoscopy is a diagnostic procedure that allows the visualization of the whole GI tract, acquiring video frames, at a rate of two frames per second, while travels through the GI tract. These frames possess rich information about the condition of the stomach and intestine mucosa. In the present paper it is compared the classification performance for small bowel tumor detection of different combinations of texture descriptors taken at different scales of the Discrete Wavelet Transform and Discrete Curvelet Transform domains. The classification step is performed by a multilayer perceptron neural network. The proposed method has been applied in real data taken from several capsule endoscopic exams and reaches 91.7% of sensitivity and 89.4% specificity for features extracted from the DWT domain and 94.1% of sensitivity and 92.4% specificity for features extracted from the DCT domain. These promising results support the feasibility of the proposed method.