METHOD FOR DIAGNOSING A PATIENT FOR SIGNS OF RESPIRATORY INFECTION BY MEANS OF CNN WITH AN ATTENTION MECHANISM AND A SYSTEM FOR ITS IMPLEMENTATION Russian patent published in 2021 - IPC B01D53/00 G16H50/00 

FIELD: information technology; medicine.

SUBSTANCE: invention relates to the field of information and communication technologies (ICT) specifically designed for medical diagnostics, in particular to a method and system for diagnosing respiratory infection based on the analysis of acoustic data of a patient using deep learning methods. A method is proposed in which, using the deep learning method, the problem of determining the probability of acquisition of a respiratory infection affecting the human respiratory tract is solved using the records of the patient’s cough, breathing and speech. In the claimed invention, a deep learning algorithm implemented in a client-server application using CNN convolutional neural networks with an attention mechanism is used to diagnose respiratory infection, the diagnosis is made based on decisions on three branches: cough, breathing and speech.

EFFECT: group of inventions is intended to provide a method and system for rapid diagnosis of respiratory infection in a patient with great accuracy.

10 cl, 9 dwg