Findings based on pulse wave dynamics, skin bloodflow patterns, and spectral characteristics, using as little as five seconds of data
TUESDAY, Sept. 8, 2026 (HealthDay News) — A machine learning algorithm can detect hypertension and diabetes from facial videos, according to a study presented at the annual meeting of the European Society of Cardiology, held from Aug. 28 to 31 in Munich.
Ryoko Uchida, from the University of Tokyo Hospital, and colleagues developed and evaluated a machine learning algorithm that detects hypertension and diabetes from a brief facial video captured by a spectroscopic camera, without physical contact or blood sampling. The analysis included more than 200 consenting participants (patients and healthy volunteers).
The researchers’ algorithm used three categories of features from the video data: pulse wave dynamics, skin blood-flow patterns, and spectral characteristics. The pulse wave-based analysis achieved 95.0 percent accuracy for detecting hypertension using a 30-second recording in an ensemble model (normal blood pressure sensitivity, 100.0 percent; hypertension sensitivity, 89.2 percent) and achieved 90.3 percent accuracy using five seconds of data (normal blood pressure sensitivity, 98.3 percent; hypertension sensitivity, 84.5 percent). Using 30-second videos of blood-flow patterns achieved 88.2 percent accuracy for diabetes detection (receiver operating characteristic-area under the curve [ROC-AUC], 0.97) and 81.2 percent accuracy using a five-second video (ROC-AUC, 0.98). For blood pressure estimation using 30-second videos of the face only, a model built on three parallel models achieved a mean absolute percentage error of 9.7 percent and a model correlation coefficient of 0.968.
“It is remarkable that AI-supported technologies are enabling the development of such powerful tools for early disease prevention,” Nico Bruining, Ph.D., editor-in-chief of the European Heart Journal – Digital Health, said in a statement. “Because this approach is quick, easy and contactless, it could be used in many settings beyond hospitals, giving it the potential to reach far more people than traditional screening methods.”
One author disclosed financial ties to the biopharmaceutical industry.
IBD Not Linked to Risk for Hypertensive Disorders of Pregnancy Overall
Mortality in Injured Children Linked to Higher SBP Thresholds Than Those That Define Hypotension
Spinal Cord Stimulation Cuts Pain, Improves Sensation With Diabetic Nerve Damage
Eating More Whole Grains Daily Linked to Cardiometabolic Benefits
Intensive BP Lowering Reduces Recurrent Stroke After Intracerebral Hemorrhage
Onset of Mortality, Morbidity Delayed for Offspring of Centenarians
Vascular Burden in Midlife Linked to Increased Risk for Dementia, Death
28 Percent of U.S. Adults With HTN Eligible for Treatment Under New Guidelines