AI in the role of emergency service: created an AI that predicts the death of firefighters
New AI model predicts sudden heart attacks in firefighters with high accuracy
Researchers at the National Institute of Standards and Technology (NIST), in association with the University of Rochester and Google, developed an artificial intelligence (AI) model that accurately predicts the occurrence of a sudden heart attack in firefighters (cause of 40% of all deaths), which is more deadly on the job than smoke and fire.
Research showed that the AI model was able to correctly identify about 97% of the abnormal electrocardiogram (ECG) samples in a unique dataset collected from 112 firefighters. If implemented, this technology could save the lives of firefighters, who are at particularly high risk of heart attacks compared to other emergency services.
Statistically, a firefighter is twice as likely to have a heart attack as a police officer, and four times more likely than other emergency services. According to experts, the limited ability to cool down in conditions of hard work and insufficient awareness of the possible risks led to such statistics.
The developed Heart Health Monitoring (H2M) model is trained to recognize and classify normal and abnormal heart rhythms indicative of disorders such as atrial fibrillation or ventricular tachycardia.
Researchers see the possibility of using the H2M model in portable heart monitors that firefighters can wear on duty. Such a device would provide real-time warnings of possible cardiac abnormalities, acting as an instant cardiologist at the scene. In addition to firefighters, this model can be useful to other high-risk groups as well as the general public if trained on appropriate datasets.
Source link
www.securitylab.ru
