Google has introduced a research system that utilizes a smartphone’s front-facing camera and artificial intelligence to estimate heart rate and resting heart rate during typical device usage. The technology, named Passive Heart Rate Monitoring During Smartphone Use in Everyday Life (PHRM), analyzes brief facial videos taken after a user unlocks their phone, measuring subtle variations in light reflected from the skin as blood circulates through facial blood vessels.
Read More
This system employs photoplethysmography, a method often found in wearable health devices, processing eight-second video clips through AI models that operate directly on the smartphone. These models provide heart rate estimates along with confidence scores that are compiled throughout the day to determine resting heart rate.
Google conducted training and evaluation of the system with over 350,000 video samples from nearly 700 participants in both structured and everyday environments, ensuring performance across a range of skin tones. A study published in Nature reported that the technology achieved heart rate measurements with an average error of less than 10% when compared to electrocardiogram readings, and the resting heart rate estimates were comparable to those from consumer fitness wearables.
In future developments, Google plans to enhance video stabilization, camera exposure, and motion detection to improve accuracy and capture measurements at optimal resting times. This research underscores the expanding potential for smartphones to act as passive health monitoring tools, minimizing the need for dedicated sensors or wearables.