Key takeaways
- This study investigated the ability of a machine-learning algorithm to detect hypertension and diabetes from facial videos.
- The algorithm detected hypertension with more than 90% accuracy and diabetes with more than 80% accuracy from 5-second videos.
- If validated, the AI model may serve as a scalable tool to enable wider screening.
Munich, Germany – 26 August 2026: Artificial intelligence (AI)-based analysis of facial video images can rapidly and accurately detect undiagnosed high blood pressure and diabetes according to a study that will be presented at ESC Congress 2026.1
Approximately 1.4 billion adults aged 30–79 years are estimated to have hypertension2 and 589 million people live with diabetes worldwide.3 Although hypertension and diabetes are among the leading modifiable risk factors for cardiovascular disease, many cases remain undiagnosed.4 Current screening relies on dedicated clinical visits or wearable devices, which limit how much of the population can be reached.
Researchers at the University of Tokyo and Institute of Science Tokyo, Japan, investigated whether AI analysis of facial videos could be used to improve diagnosis of hypertension and diabetes. Presenter, Ms Ryoko Uchida, explained: "We aimed to develop an AI algorithm that enables contactless screening in everyday environments to detect common conditions earlier and at scale."
A prospective single-centre study recruited 215 participants, involving both diagnosed patients and healthy volunteers.5Each participant underwent a short, high-speed video recording of their face and palms using a spectroscopic camera. A machine-learning algorithm analysed the individual videos and extracted data on pulse-wave dynamics (which measure the stiffness of arteries), skin blood-flow patterns and the spectral characteristics of skin colouring. Participants also had conventional assessments to identify hypertension and diabetes.
As published previously, the algorithm was able to detect hypertension with 95.0% accuracy from a 30-second recording using pulse wave-based analysis of both facial and palm video.5 The sensitivity to detect normal blood pressure was 100.0%, while hypertension sensitivity was 89.2%. Accuracy was still high – 90.3% – using a 5-second video.
In new research, based on facial blood flow patterns, the algorithm was able to detect diabetes with similarly high accuracy: 88.2% from a 30-second video and 81.2% from a 5-second video.
The algorithm could also estimate blood pressure from a facial video alone, without a cuff. The mean absolute percentage error for systolic blood pressure (SBP) was 8.6%. The algorithm had a mean error of −2.6 mmHg for SBP, within the Association for the Advancement of Medical Instrumentation (AAMI)'s limit of ±5.0 mmHg, although the standard deviation error was ±12.0 mmHg, exceeding the AAMI criterion of ±8.0 mmHg. Future work will focus on reducing this variability through larger, multicentre datasets and feature optimisation.
Ms Uchida concluded, "Our machine-learning algorithm accurately detected hypertension and diabetes from facial spectroscopic video recordings as short as 5 seconds. We intend to validate these findings in larger cohorts across more diverse populations to support real-world application. If validated, this contactless approach could allow people to be screened in everyday settings – without cuffs, blood sampling or a dedicated clinic visit – helping to identify at-risk individuals who would otherwise remain undiagnosed and therefore untreated."
"It is remarkable that AI-supported technologies are enabling the development of such powerful tools for early disease prevention," commented Associate Professor Nico Bruining, Programme Co-Chair of the ESC Digital and AI Summit and the Editor-in-Chief of the European Heart Journal – Digital Health. "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. Detecting these conditions early means treatment and lifestyle changes can start sooner, helping to prevent heart attacks, strokes and other cardiovascular diseases."
The latest advances in AI and cardiovascular care will be discussed at the ESC Digital & AI Summit in Basel, Switzerland, on 12–13 November 2026. The event provides a deep dive into many other exciting AI innovations that are transforming cardiovascular care and prevention. The summit will be attended by the rapidly growing community of clinicians, researchers, innovators and industry experts who are shaping the future of cardiovascular health.
