Scientists apply deep learning to blood pressure measurements
Deep learning, a process that continuously analyzes data, is now being incorporated into medical devices.
In a recent development, the technology was applied to blood pressure measurements, with promising results.
Park Se-young has the details.
A sphygmomanometer, or blood pressure gauge, is probably one of the most common ways to measure blood pressure, but it's generally less accurate than the readings doctors get using a blood pressure pump and a stethoscope.
However, that may not be true for long.
A team of Korean researchers has found a way to improve accuracy through the application of deep learning technology.
The researchers created a system similar to AlphaGo, a computer program that uses artificial intelligence and deep learning to play baduk.
It studied thousands of baduk matches, and eventually beat grandmaster Lee Sedol -- a feat previously thought impossible.
Using similar techniques, the researchers fed the blood pressure values of 85 people taken by a sphygmomanometer and 85 taken by a doctor... into their system.
After studying the data, the system was able to match the values attained by doctors and produce similar results.
It achieved an accuracy rate of 95 percent, which was an improvement from the 90-percent accuracy of the blood pressure gauges.
"By studying the collected data, the system was able to provide more accurate measurements."
The researchers hope their finding will eventually be transformed into a new diagnostic tool for use at home and in hospitals.