k-perry/pulse_observer

Detects a person's heart rate via webcam by tracking tiny changes in face colorization due to blood flow.

Pulse Observer

Description

Compute a person's pulse via webcam in real-time by tracking tiny changes in face coloration due to blood flow. For best results, try to minimize movement.

Uses Python, OpenCV, NumPy, SciPy, and Dlib.

Requirements

• Python 2.7.x
• OpenCV 3.0
• NumPy 1.11
• SciPy 0.17.1
• Dlib 18.17.100
• Download Dlib's pre-trained predictor model for facial landmarks and put it in the same directory as pulse_observer.py.

Compatibility

• This program has only been tested on Windows 10. It may or may not work on Linux or Mac OS.
• It will only work with Python 2.7.x. It will not work with Python 3+.
• Tested with the following webcams:
  • Logitech C270
  • Logitech C920
  • Asus laptop

Author

Kevin Perry (kevinperry@gatech.edu)