Identification of SARS-CoV-2 Pneumonia in Chest X-ray Images Using Convolutional Neural Networks

In 2019, COVID-19 disease emerged in Wuhan, China, leading to a pandemic that saturated health systems, raising the need to develop effective diagnostic methods. This work presents an approach based on artificial intelligence applied to X-ray images obtained from Mexican patients, provided by Hospital General de Zona No. 24. A dataset of 612 images with 2 classes: COVID and HEALTHY, were labelled by a radiologist and also verified with positive RT-PCR test. The first class contains X-ray images of patients with pneumonia due to SARS-CoV-2 and the second contains patients without diseases affecting the lung parenchyma. The proposed work aims to classify COVID-19 pneumonia using convolutional neural networks to provide the physician with a suggestive diagnosis. Images were automatically trimmed and then transfer learning was applied to VGG-16 and ResNet-50 models, which were trained and tested using the generated dataset, both achieving an accuracy, recall, specificity and F1-score of over 98%. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.

Linear Sintonization of Two Cascade Vectorial Controllers from the Numerical Prediction of a Mechanical Ventilator System

With the increasing number of COVID-19 cases, the demand of ventilators has increased to such an extent that there is not enough supply of equipment available. For this reason, this research is based on an open source project presented by the Massachusetts Institute of Technology which involves the conditioning of an Artificial Manual Breathing Unit (AMBU bag), that acts as an aid for patients to maintain a constant respiratory cycle. This document presents the calculation of the control matrix for the mechatronic and bio pneumatic systems of the ventilator by implementing a Space-State Model that allows to obtain the control gains that may reduce the error concerning the open loop operation of the ventilator;furthermore, this paper also presents the SIMULINK simulations that demonstrate the efficiency of the calculated control matrixes. Due to these graphic demonstrations, it was observed that the implemented controllers provided an increase in the response time and reduced the error, thus, theoretically, both controllers must provide stability and a higher performance to the ventilator prompting an appropriate assistance to the hospitalized victims of the COVID-19;however, currently, this project is not viable, as it has not been tested in real life and it does not meet the quality and medical requirements for its use in emergencies. © 2020 IEEE.