Artificial Classification System for Urothelial Carcinoma

Abstract

This paper presents an artificial classification system (ACUC) that can be applied to cases of urothelial carcinoma. The ACUC was combined with a microscopy system to enable cell images to be captured from slides and subsequently transferred to a computer for classification. We introduce a two-stage convolutional neural network (CNN) model to classify high-grade urothelial carcinoma. The complexity of the CNN architecture can be reduced using a single CNN model. The ACUC was tested on 600 segments of cell sample images, which were provided by the E-DA hospital, and the results indicated that the accuracy of the ACUC is approximately 88%.