Non-Contact Heart Rate Measurement Using Commodity Cameras and AI

Traditional heart rate measurement typically relies on contact-based devices, such as medical equipment or wearable sensors. This paper introduces a novel system that enables non-contact, real-time heart rate measurement using readily available commodity cameras, like those embedded in laptops. The system employs an innovative algorithm to capture relevant signals from video streams in real-life environments. The heart rate computation process involves four main steps: identifying camera frames per second, performing face detection with 68 facial landmarks using deep learning, applying a time sliding window algorithm for signal denoising, and finally computing heart rate based on identified signal periodicity. Prototypes were tested against Apple Watch measurements, showing comparable results. The researchers plan further tuning and optimization, with the ultimate goal of deploying the system as a personal AI agent for continuous health monitoring, particularly beneficial for elderly care.