Researchers from the Kulakov National Medical Research Center and RTU MIREA have developed and piloted an AI-based service for analyzing cardiotocography (CTG) data. The system acts as a digital assistant, providing physicians with a second opinion to help predict perinatal outcomes during labor.
The technology allows doctors to capture images of CTG paper strips or monitor screens using a smartphone. The AI processes these images in seconds to identify potential fetal hypoxia, which is critical for making rapid clinical decisions, such as performing an emergency cesarean section.
The service utilizes the secure Russian platform 'Max' for image transmission. The underlying neural network employs transformer architectures to segment, normalize, and classify CTG data, effectively compensating for distortions in photographs of physical records.
During the evaluation phase, the AI solution demonstrated an 86% accuracy rate in predicting perinatal outcomes. The project aims to standardize diagnostic data and improve access to expert-level analysis across medical facilities in Russia.
