Chest x-ray

AI solution that identifies abnormalities on chest x-rays for faster, more accurate diagnosis

How does it work?

The algorithm examines images and, based on a special protocol, classifies them into two categories:

• with identified pathological changes

• without identified pathologies

 

Based on the results, the system generates a conclusion and provides a research protocol. The algorithm detects any signs of pathologies, including oncology, pneumonia and tuberculosis.

How the service works

Chest x-ray is based on InсeptionV3, Inсeption ResNet-2, ResNet-50 neural networks. They use image recognition to mark samples up according to a certain algorithm, learning on large datasets from x-ray and medical reports. Neural networks are supplemented with a specific environment for their interaction with users. 

Model accuracy > 81%
Patient admission speed increases by 30%

The service provides

Support in making clinical decisions by sorting and prioritizing research of pathologies
Reduced burden on doctors due to automatic detection of affected areas
Improved early diagnosis, including serious diseases
Reduced risk of medical errors

Registration

Registration certificate of medical devices dated December 11, 2020 No. RZN 2020/11137