What is it about?
We compare three technologies for measuring breathing and heart rate without attaching sensors to the body: radio-frequency (RF) sensing, lidar, and cameras. All three systems observed the same people at the same time, while wearable sensors provided reference measurements for comparison. We also introduce an AI-assisted camera method that estimates breathing from small changes in chest depth using ordinary RGB video, without requiring a physical depth camera or specialized training. Our results show that each sensing technology has different strengths and limitations depending on factors such as body position, orientation, and movement.
Featured Image
Photo by engin akyurt on Unsplash
Why is it important?
Contactless monitoring of breathing and heart rate could support applications such as sleep monitoring, assisted living, patient observation, and smart-home health monitoring. Our results show that no single sensing technology performs best in every situation. Cameras and lidar performed particularly well for breathing measurements, while RF achieved the highest heart-rate detection probability. Because the sensors tend to fail under different conditions, combining them could enable more reliable and robust vital-sign monitoring than relying on any one technology alone.
Perspectives
A key finding of this work is that RF, lidar, and cameras provide complementary information rather than simply competing ways of measuring the same signal. When one sensing method struggles because of body orientation, movement, visibility, or other conditions, another may still provide useful physiological information. This creates an opportunity for future multimodal sensing systems to automatically determine which sensor, or combination of sensors, is most reliable at a given moment.
Dr. Raied Caromi
National Institute of Standards and Technology
Read the Original
This page is a summary of: Cardiorespiratory Monitoring using RF, Lidar, and Camera: A Comparison against Synchronized Multimodal Measurements, October 2026, ACM (Association for Computing Machinery),
DOI: 10.1145/3842657.3842786.
You can read the full text:
Contributors
The following have contributed to this page
