Webcam heart rate monitoring rivals Apple Watch accuracy

Researchers Kelly Li and Fulu Li have introduced a non-contact, real-time heart rate measurement system using commodity cameras, such as a laptop's built-in webcam. The method processes video feeds at 30 frames per second, first detecting faces via deep learning and identifying 68 facial landmarks. A time sliding window algorithm then smooths out noise to extract periodic signals correlated with heartbeats, enabling accurate computation of heart rate.

The system was tested against an Apple Watch, with multiple rounds of measurement on the same person at the same time. The difference range was reported, and the mean difference calculated, demonstrating strong correlation. The authors plan further optimization and deployment as a personal AI agent for continuous health monitoring, particularly for elderly individuals who may not wear smartwatches or traditional sensors.