Photoplethysmograph (PPG) based Biometric Recognition

  • Amongst all medical biometric traits, Photoplethysmograph (PPG) is the easiest to acquire, as it doesn't require gel, external stimulus or multiple electrodes.
  • PPG records the blood volume change with just combination of Light Emitting Diode and Photodiode from any part of the body. With IoT and smart homes' penetration, PPG recording can easily be integrated with other vital wearable devices.
  • PPG being a physiological signal, it is harder to steal or replicate. It has advantages of inherent anti-spoofing and liveness detection over traditional biometrics modalities.
  • PPG represents peculiarity of hemodynamics and cardiovascular system for each individual. Our research aims to leverage this for biometric recognition.

    • ppg      

    BioSec.Lab PPG Dataset (Biosec1)

    The BioSec.Lab PPG dataset (Biosec1) was created for research on Photoplethysmograph (PPG) based biometrics recognition at University of Toronto.

    This dataset includes signals recorded in four different conditions to evaluate permanence, robustness and uniqueness of PPG signal as a biometric identity. Dataset contains signals in following settings:

    1. Relax condition: 3 minutes long PPG signal recorded from 86 subjects.
    2. After Exercise: 3 minutes long PPG signal recorded from 40 subjects after heavy exercise.
    3. Short time-lapse: 3 minutes each PPG signal recorded in two parts at least 30 minutes apart on same day from 55 subjects.
    4. Long time-lapse: 3 minutes each PPG signals recorded in two parts at least 2 weeks apart from 37 subjects.

    In addition, we also have,

    • Fingertip video from mobile camera: 1 minutes long video recording with fingertip on camera lens from 36 subjects.

    Read more...

    How to use

    This dataset is available for research purposes. More details on recording conditions and device are provided on the dataset description page. Dataset is arranged in .mat files for all settings.

    To access files, contact Prof. Hatzinakos or Dae Yon Hwang. Dataset would be provided after authorizations and End User License Agreement.

    Publication

    Evaluation on this dataset was done using several different methods in verification mode in following paper:

    • U. Yadav, S.N.Abbas, D. Hatzinakos, "Evaluation of PPG Biometrics for Authentication in different states" The 11th IAPR International Conference on Biometrics (ICB), Goldcoast, Australia, Feb 2018. ( Preprint )

    Credits

    We'd like to thank all participants of this study. We'd also like to thank NSERC and Mitacs for their support.