Vision-Based Cardiorespiratory Monitoring
AssigneeMississippi State University
InventorAmirtaha Taebi
Life of this patent
Estimating the term…
Abstract
A system and method for non-invasive cardiorespiratory health assessment using a camera to capture video of a subject's chest region. The method includes tracking the motion of pixel groups in the video to construct a 3D motion map of the chest surface vibrations. The motion map is then analyzed by an artificial intelligence (AI) model to classify the subject's cardiorespiratory health status and estimate various physiological parameters. The AI model is trained on a dataset of chest vibration maps labeled with corresponding health statuses. The system can be implemented on various devices, including smartphones, and can provide real-time feedback to the user through a user interface.
Claims20
A method, comprising:
- capturing a video of a chest region of a subject using a camera, wherein the video comprises a plurality of consecutive frames;
- for each frame in the video, tracking a plurality of pixel groups within the frame;
- determining a movement of each of the plurality of pixel groups between consecutive frames;
- and
- constructing a motion map of the chest region based on the movements of the plurality of pixel groups, wherein the motion map represents vibrations of the chest region.
A method, comprising:
- capturing a video of a chest region of a subject using a camera;
- generating a chest motion map based on the captured video;
- inputting the chest motion map into an artificial intelligence (AI) model for classifying cardiorespiratory health;
- and
- classifying cardiorespiratory health of the subject into a predefined cardiorespiratory category based on a classification result from the AI model, wherein the classification result indicates cardiorespiratory health status of the subject.
An apparatus, comprising:
- a camera configured to capture a video of a chest region of a subject;
- and
- a processing unit coupled to the camera, the processing unit configured to:
- track a plurality of pixel groups within each frame of the video;
- determine a movement of each of the plurality of pixel groups between consecutive frames;
- construct a motion map of the chest region based on the movements of the plurality of pixel groups, wherein the motion map represents vibrations of the chest region in three dimensions (3D);
- generate a chest vibration map based on the motion map;
- and
- input the chest vibration map into an AI model for cardiorespiratory health assessment.
Dependencies are inferred from the claim text.