Development of a Neurocomputer Modular Information System for Cancerous Diseases Diagnostics in Animals

Abstract

Information system was developed in the form of a web application that makes it possible to identify microscopic images of cytological samples, to establish an initial diagnosis and to provide recommendations for its further confirmation based on additional data. Approaches, assumptions and prerequisites adopted in the information system development are described. It is proposed to use neural networks as the information system element in sample identification and making the initial diagnosis. Patient data, affected area images and microscopic images of cytological samples are planned to be collected in the information system database under creation. Cytological sample images serve as the input data for neural networks operation. Cytological picture assessment is based on the use of the following characteristic features: preparation background, number and location of cells, size and shape of cells, nucleus, presence of multinucleated cells and fission entities (atypical mitoses), etc.