Heart Rate Detection Using Microsoft Kinect: Validation and Comparison to Wearable Devices.

Area of Science:

• Biomedical Engineering
• Healthcare Technology
• Signal Processing

Background:

• Contactless detection is an emerging healthcare innovation for unobtrusive biomedical parameter measurement.
• Traditional heart rate (HR) monitoring devices like ECG and pulse oximeters can be expensive or uncomfortable.
• Contactless HR detection enables continuous, device-free monitoring and clinical support.

Purpose of the Study:

• To validate a contactless HR estimation method using RGB data from a Microsoft Kinect v2 device.
• To compare the performance of the contactless method against gold-standard (Holter) and commercial smartwatch devices.
• To assess the impact of prior knowledge about subjects' HR on estimation accuracy.

Main Methods:

• Utilized Eulerian Video Magnification (EVM), Photoplethysmography (PPG), and Videoplethysmography (VPG) on RGB data.
• Employed a Holter monitor as the ground-truth for HR measurement.
• Conducted validation in two modalities: with and without a priori knowledge of subjects' HR.

Main Results:

• The contactless HR estimation method achieved performance comparable to wearable systems under specific conditions.
• HR estimation accuracy was 2% deviation from ground-truth when subject HR information was known.
• HR estimation accuracy showed a 3.4% deviation from ground-truth when no prior subject information was available.

Conclusions:

• Contactless HR detection using RGB video is a viable, non-invasive alternative to traditional methods.
• The method demonstrates high accuracy, particularly when integrated with prior subject knowledge.
• This technology holds promise for remote and continuous patient monitoring in healthcare settings.