Microsoft Kinect Visual and Depth Sensors for Breathing and Heart Rate Analysis

Area of Science:

• Biomedical Engineering
• Computational Intelligence
• Medical Diagnostics

Background:

• Non-contact monitoring of physiological signals is crucial for early detection of medical and neurological disorders.
• Existing methods often require specialized equipment or direct physical contact, limiting their applicability.
• The Microsoft (MS) Kinect sensor offers a potential solution for accessible, non-invasive vital sign monitoring.

Purpose of the Study:

• To introduce and validate a novel method for non-contact breathing and heart rate estimation using MS Kinect sensors.
• To assess the accuracy of this method in detecting biomedical features relevant to medical and neurological conditions.
• To demonstrate the potential for using this technology in home environments or during physical activities.

Main Methods:

• Utilizing MS Kinect's image, depth, and infrared sensors to record facial and thorax movements.
• Applying computational methods, including data selection, denoising, spectral analysis, and visualization.
• Analyzing time-series data from selected regions of interest, particularly the mouth area and thorax.
• Comparing results with contact-based measurements from Garmin sensors.

Main Results:

• Verified correspondence between breathing rate estimations from image/infrared data (mouth) and depth data (thorax).
• Successfully estimated heart rate using spectral analysis of mouth area video frames.
• Achieved high accuracy: 0.26% for non-contact breathing rate and 1.47% for heart rate estimation (infrared sensor).
• Demonstrated the ability to differentiate breathing types using video frames and depth data.

Conclusions:

• MS Kinect sensors can efficiently capture multidimensional biomedical data for detecting specific features.
• Computational intelligence methods enable accurate non-contact vital sign monitoring.
• The proposed method supports diagnostic purposes in diverse settings, enhancing human-machine interaction for healthcare.