NEWS
中文Research Highlights
- Publish Date：2025-09-18
NYCU Develops Smartphone-Based, Contactless System for Heart Rhythm Monitoring Without ECG
The contactless heart monitoring technology developed by NYCU was showcased at CES in the United States.
Edited by Chance Lai
______
______
What if checking your heart health was as easy as looking into your phone’s camera? A research team led by Professor Bing-Fei Wu at the Institute of Electrical and Control Engineering, National Yang Ming Chiao Tung University (NYCU), has developed a breakthrough system that can detect atrial fibrillation (AF)—a significant risk factor for stroke—using only the camera of a smartphone or laptop.
This non-contact, lightweight solution enables users to monitor heart rhythms in real-world settings, without the need for traditional ECG devices or physical sensors.
Atrial Fibrillation, Reimagined for Everyday Life
AF is closely associated with stroke risk, yet it often goes undetected until it’s too late. Conventional detection methods rely heavily on contact-based equipment, such as ECGs, which can be uncomfortable to wear for extended periods and are not always accessible outside of clinical settings.
To address this critical gap, Prof. Wu’s team turned to remote photoplethysmography (rPPG)—a technique that captures microvascular color changes on a person’s face via a standard camera. By analyzing these subtle signals, the system accurately estimates heart rate data in real-time.
Smart AI, No Cloud Required
The team also introduced a novel signal processing algorithm that significantly reduces interference caused by head movement and lighting changes—two common challenges in daily environments. Instead of relying on computationally intensive deep-learning models, the system employs a lightweight AI architecture with significantly reduced parameters and minimal latency.
This means it can deliver high-performance analysis without an internet connection, opening new frontiers in offline, personalized health monitoring.
Clinically Validated with 450+ Subjects
To ensure clinical reliability, the team partnered with Dr. Yu Sun from En Chu Kong Hospital to establish a comprehensive video database featuring over 450 volunteers. The dataset includes recordings of individuals with normal heart rhythms, AF, and other arrhythmias, captured under realistic lighting and motion conditions.
Even in these challenging environments, the system demonstrated high accuracy and stability, earning recognition from both the academic and tech communities.
This non-contact, lightweight solution enables users to monitor heart rhythms in real-world settings, without the need for traditional ECG devices or physical sensors.
Atrial Fibrillation, Reimagined for Everyday Life
AF is closely associated with stroke risk, yet it often goes undetected until it’s too late. Conventional detection methods rely heavily on contact-based equipment, such as ECGs, which can be uncomfortable to wear for extended periods and are not always accessible outside of clinical settings.
To address this critical gap, Prof. Wu’s team turned to remote photoplethysmography (rPPG)—a technique that captures microvascular color changes on a person’s face via a standard camera. By analyzing these subtle signals, the system accurately estimates heart rate data in real-time.
Smart AI, No Cloud Required
The team also introduced a novel signal processing algorithm that significantly reduces interference caused by head movement and lighting changes—two common challenges in daily environments. Instead of relying on computationally intensive deep-learning models, the system employs a lightweight AI architecture with significantly reduced parameters and minimal latency.
This means it can deliver high-performance analysis without an internet connection, opening new frontiers in offline, personalized health monitoring.
Clinically Validated with 450+ Subjects
To ensure clinical reliability, the team partnered with Dr. Yu Sun from En Chu Kong Hospital to establish a comprehensive video database featuring over 450 volunteers. The dataset includes recordings of individuals with normal heart rhythms, AF, and other arrhythmias, captured under realistic lighting and motion conditions.
Even in these challenging environments, the system demonstrated high accuracy and stability, earning recognition from both the academic and tech communities.
Global Recognition and Real-World Application
The research results were published in the IEEE Journal of Biomedical and Health Informatics, where the study was selected as a Feature Article. The project also won the Excellence Award in Artificial Intelligence at the 2024 TSC Thesis Awards (崇越論文大賞).
Most notably, the technology was deployed in commercial devices, such as laptops and smartphones, and showcased in the FaceHeart CardioMirror. This intelligent health mirror won a CES 2025 Innovation Award in Digital Health at the world’s largest consumer tech event.
A Game-Changer for Telehealth and Preventive Care
This innovation isn’t just a lab prototype—it’s a real-world solution with the potential to transform telemedicine, community screening, and early diagnosis for high-risk groups. It empowers individuals to detect signs of cardiovascular distress early, giving doctors and patients more time to act before emergencies strike.
As the world continues to shift toward remote healthcare, NYCU’s contactless AF monitoring system exemplifies the power of human-centered AI to make everyday health smarter, safer, and more accessible.
Prof. Bing-Fei Wu, Institute of Electrical and Control Engineering at NYCU (Photo credit: Far Eastern Y.Z. Hsu Foundation)Related Image(s)：
