Utilization of Convolutional Neural Networks for Urinary Bladder Cancer Diagnosis Recognition From CT Imagery

Abstract

In this paper, an approach for urinary bladder cancer diagnosis from computer tomography (CT) images based on the application of convolutional neural networks (CNN) is presented. The image data set that consists of three main parts (frontal, horizontal, and sagittal plane) is used. In order to classify images, pre-defined CNN architectures are used. CNN performances are evaluated by using 5-fold cross-validation procedure that gives information about classification and generalization performances. From the presented results, it can be noticed that higher performances are achieved if more complex CNN architectures are used. Higher performances can be noticed regardless of a plane in which images are captured. An increase in performances can be noticed in both classification and generalization context.