Authors say technology may enable noninvasive screening
By Lori Solomon HealthDay Reporter
MONDAY, Nov. 18, 2024 (HealthDay News) — Machine learning analysis of photo and video data demonstrates a promising way to noninvasively detect early hypertension and diabetes, according to a study presented at the American Heart Association Scientific Sessions 2024, held from Nov. 16 to 18 in Chicago.
Ryoko Uchida, from University of Tokyo, and colleagues developed a machine learning algorithm to analyze skin blood flow and spectral features from photos to noninvasively assess biometric information for health monitoring. The analysis included images from approximately 200 participants’ faces and hands, as well as blood pressure readings.
The researchers reported that the algorithm detected hypertension with 94.2 percent alignment with the American Heart Association guidelines for stage 1 hypertension, based on pulse wave features. Among 40 percent who had early hypertension, accuracy rates were 86.2 percent for 30-second data segments and 80.9 percent for 5-second segments. When applying the algorithm to video data, it achieved a 75.3 percent accuracy in detecting diabetes, utilizing blood flow patterns as markers.
“This method may someday allow people to monitor their own health at home and could lead to early detection and treatment of high blood pressure and diabetes in people who avoid medical exams and blood tests,” Uchida said in a statement.
Copyright © 2024 HealthDay. All rights reserved.
