And just as AlphaGo studied thousands of baduk matches, the team inputted blood pressure values – from a sphygmomanometer and a doctor – into its AI system for deep learning. (image: KobizMedia/ Korea Bizwire)
SEOUL, Nov. 8 (Korea Bizwire) – A sphygmomanometer is a quick and convenient way of measuring blood pressure. But it tends to lack accuracy when compared to measurements made by a doctor with a blood pressure pump and a stethoscope.
To overcome these limits, a team of Korean scientists from Hanyang University applied deep learning technology to an AI system for blood pressure measurements, successfully improving its accuracy.
Deep learning is a learning process of a machine or an artificial intelligence that continuously analyzes data to boost its performance, ideally to the point where it can develop the capacity to make decisions and draw conclusions. The concept was made popular in Korea earlier this year when Google’s deep learning AI AlphaGo challenged Lee Se-dol in a five-game baduk match.
And just as AlphaGo studied thousands of baduk matches, the team inputted blood pressure values – from a sphygmomanometer and a doctor – into its AI system for deep learning.
After numerous self-studies, the system was able to provide a doctor’s value, upon the entry of data from a blood pressure meter, with a 95 percent accuracy, which was an improvement from an average of 90-percent accuracy conveyed by blood pressure meters.
“We hope our research result finds its place as a new bio-diagnostics method with a widespread application at homes and in hospitals,” said Dr. Chang Joon-hyuk, who led the research.
The full study findings were published in the October 15 edition of Transactions of Industrial Informatics.
By Lina Jang (linajang@koreabizwire.com)
