Convolutional Neural Networks in the Ovarian Cancer Detection

The mainKowalski, Piotr A. goalBłoniarz, Jakub ofChmura, Łukasz the work is to create an artificial neural network capable of recognizing tumor cell groups in histopathological slides. Here, the Convolution Neural Network was used in the course of research on the detection of areas that include ovarian cancer cells. Theoretical and practical investigations included the issues of both preparing data for learning and selecting the appropriate architecture of the artificial neural network. The obtained results show that neural networks can be used as a competitive method for designating areas in which pathological cells of cancer origin are present. An additional goal of the work is to obtain superior results without losing the quality of the input image as one of the major limitations to this is the memory and computing capabilities of the computer on which the tests are carried out.