Hosted by IDTechEx
Artificial Intelligence Research
Posted on January 15, 2018 by

AI detects cardiac arrest

Emergency services in Denmark are enlisting the help of artificial intelligence to analyse emergency calls along with a human operator.
 
Developed by start-up Corti, the system listens in for signs of a possible cardiac arrest when a call is made to 112 in Copenhagen. The AI system uses speech-recognition software to transcribe what is being said, confirms a diagnosis and then flashes an alert on the screen for the operator to see.
 
Conversations are noisy, implicit, and hard to understand, but they contain a goldmine of information according to Corti who have developed a multitude of deep neural networks that listen directly to a sound stream and extract the most important features. The better the quality of these features, the better their prediction and reasoning frameworks are.
 
The AI system creates incredibly fast and accurate predictions by combining machine learning models that utilize advanced feature extraction capabilities. By applying convolutional and recurrent neural networks, the AI examines a larger contextual input from which the models train enabling even subtle correlations in data with high accuracy.
 
When a patient or a bystander makes an emergency call, the dispatcher at the other end of the line must triage the patient on their own. With Corti implemented, the dispatcher now gets a digital assistant that listens in on the conversation and helps to look for important signals in both verbal communication as well as tone of voice and breathing patterns, while also considering other metadata.
 
 
All the data provided during the emergency call is automatically analyzed by Corti and then compared to the millions of emergency calls, which Corti has already analyzed, to find important patterns.
 
As Corti's understanding of the incident increases, Corti will try to predict the criticality of the patient's situation based on symptom descriptions and the signals gathered from voice and audio. These insights are delivered to the dispatcher as alerts and recommendations.
 
"This is an innovation with the potential to change the way Emergency Medical Services handle emergency calls." says Freddy Lippert, MD CEO EMS Copenhagen.
 
 
Source and top image: Corti
More IDTechEx Journals