EUVIMEDSkopri EUVIMED
Open access

Deep learning–integrated multilayer thermal gradient sensing platform for real-time blood flow monitoring

Youngmin Sim iD, Yosep Park iD, Kyeongha Kwon iD

Science Advances · 2026-02-06 · 12 (6)

DOI: 10.1126/sciadv.aea8902 ↗

At a glance

Journal article in Science Advances, published on 2026-02-06. Written by Youngmin Sim and 2 co-authors. Institution: School of Electrical Engineering, Korea Advanced Institute of Science and Technology, Daejeon, Republic of Korea. Major MEDLINE topics: Deep Learning. Cited 2 times so far (Crossref); 36 references. The full text is freely available in PubMed Central. Funded by: Samsung Advanced Institute of Technology, National Research Foundation of Korea, BK21 FOUR program.

AstrattSubjectsDettalji biblijografiċiCiteRelated studies

Abstract (original language)

Blood flow monitoring is fundamental for assessing cardiovascular health and identifying vascular complications. Traditional Doppler ultrasound methods require bulky equipment and specialized expertise, while recent thermal sensing approaches face limitations due to blood vessel depth variability beneath the skin. We present a soft electronic platform that integrates multilayer thermal sensing with deep learning algorithms to simultaneously measure blood flow rate and vessel depth. The device uses a wireless system with thermal sensing modules, featuring strategically positioned thermistors in separate layers to capture thermal gradients at different heights from the skin surface. Deep learning processes multilayer thermal patterns to extract both parameters in real time. Validation through benchtop testing, finite element analysis, and on-body trials demonstrates accurate measurements across relevant flow rates and vessel depths. Integration with photoplethysmography enhances continuous blood pressure monitoring accuracy compared to conventional approaches, particularly during dynamic physiological changes. This technology offers potential for personalized cardiovascular monitoring, early detection of hemodynamic events, and skin graft surveillance.

Subjects

MeSH terms (MEDLINE)

Deep LearningHumansPhotoplethysmographyMonitoring, PhysiologicAlgorithms

Authors and institutions

  • School of Electrical Engineering, Korea Advanced Institute of Science and Technology, Daejeon, Republic of Korea.
  • Graduate School of Artificial Intelligence Semiconductor, Korea Advanced Institute of Science and Technology, Daejeon, Republic of Korea.

ORCID profiles: Youngmin Sim, Yosep Park, Kyeongha Kwon

Funding

  • Samsung Advanced Institute of Technology
  • National Research Foundation of Korea
  • BK21 FOUR program
  • ICT & Future Planning
  • Institute of Information & communications Technology Planning & Evaluation (IITP) under the Graduate School of Artificial Intelligence Semiconductor

Cite

APA

Sim, Y., Park, Y., & Kwon, K. (2026). Deep learning–integrated multilayer thermal gradient sensing platform for real-time blood flow monitoring. Science Advances, 12(6). https://doi.org/10.1126/sciadv.aea8902

Vancouver

Sim Y, Park Y, Kwon K. Deep learning–integrated multilayer thermal gradient sensing platform for real-time blood flow monitoring. Science Advances. 2026;12(6). doi:10.1126/sciadv.aea8902

BibTeX
@article{Sim2026,
  author = {Sim, Youngmin and Park, Yosep and Kwon, Kyeongha},
  title = {Deep learning–integrated multilayer thermal gradient sensing platform for real-time blood flow monitoring},
  journal = {Science Advances},
  year = {2026},
  volume = {12},
  number = {6},
  doi = {10.1126/sciadv.aea8902}
}
Start for free →