Development of an Automated Medical Information System for the Analysis of Fluorograms

This article highlights aspects of the development of a medical information system for the analysis of fluorograms using convolutional neural networks. As part of the work done, an algorithm for the operation of the module for recognizing positive and negative images, as well as the implementation of this algorithm, was obtained. The developed module is of practical value for organizing a patient’s medical profile and reduces the human factor in the process of processing medical images. It is assumed that the developed system will have the required level of process automation and will consist of several interconnected blocks, including special software, the ability to send processed data via the TCP/IP protocol to the ports of the router, as well as to confirm that the received image corresponds to a standard healthy patient, or to detect deviations. The processed version of the image is sent back to the institution where the patient is being examined.