News

Care Mentor AI: Progress in the Development of a Neural Network for COVID-19 Detection in CT Studies


Care Mentor AI: Progress in the Development of a Neural Network for COVID-19 Detection in CT Studies

Care Mentor AI creates a machine vision product based on neural networks that will help doctors recognize pathologies in CT studies. The program visualizes pathologically altered regions of the lungs on axial sections of a native CT scan. With high predictive accuracy, Care Mentor AI determines the radiological signs of coronavirus infection of the lungs (COVID-19). We obtained a basic model for segmentation of CT sections to identify pathological changes of coronavirus.

Computer vision model Care Mentor AI shows excellent results on pathologies such as infiltration, however, to obtain similar results on hydrothorax requires more training of the model. This is due to lack of data in the training dataset. To improve the results and speedy development of the neural network, we need a large amount of data: CT images of patients with coronavirus infection, as well as a description of the process and progress of treatment. It takes only about a month to finalize the algorithm, and we will be able to implement the product in medical institutions. It will facilitate the work of doctors and assist them in the rapid detection of cases with pathologies.

We ask you to assist us in obtaining studies of CT of the chest organs and overview radiographs with cases of confirmed diagnosis of COVID-19.

Why CT? This week, 45 CT centers based on clinics for early diagnosis of coronavirus were launched in Moscow. This is connected with the fact that, according to A. Rakova, the deputy mayor of Moscow for social development: “it is not always possible to make a diagnosis of coronavirus infection based on laboratory tests only. Therefore, we decided to create special centers where patients with suspected coronavirus will undergo additional CT and other studies to make an accurate diagnosis.”

For all matters of cooperation - follow the link and leave your contacts, we will contact you on the same day:  carementor.me/contact