Application of artificial neural networks for classification of colonoscopic images
A Min Tjoa, Shankar Muthu Krishnan, C.J. Yap, Sandhiya Swaminathan, Peicheng Wang · Asia Pacific Conference on Circuits and Systems · 2003
Computer-aided methods in colonoscopy play an important role of assisting the physician in detecting abnormalities by characterizing the features of the segmented colonoscopic images. A novel image processing approach is proposed to extract the parametric features of the colonoscopic image. A comparison between two different types of neural networks, viz., the backpropagation (BP) and the adaptive resonance theory (ART) networks, is carried out to classify the status of the colon. The preliminary results obtained by the proposed methods are encouraging.