Welcome to the rPPG benchmark webpage!


This webpage provides an overview of scientific publications of rPPG algorithms and software.
The tables below report rPPG's test outcomes on its accuracy in extracting heart rates from skin tissue that were video recorded with consumer-level cameras.

Have you developed new rPPG software and would you like to assess the accuracy of your algorithm?
Then send your software program or MatLab script with explanations to the cripted e-mailaddress below.

In case you have also a video database available for rPPG benchmarking, we would appreciate it if you could share your database with us.
Your database will be used to benchmark the accuracy of existing software published on this webpage.
To share your database, please contact me with the e-mailaddress below and provide me a download link to the videos and a reference of a publication in which the content of the database is explained in detail.
This publication will be cited on the current webpage.

Contact details:

mygivennamemysurname[at]gmail.com

Marnix Naber
Experimental Psychology
Utrecht University
The Netherlands

rPPG tested on van der Kooij & Naber (2009) 30 second video recordings of faces, wrists, and calfs, during rest or after physical exercise.
Camera: Microsoft Lifecam HD-3000; Resolution: 1280x720 pixels; Frames per second: 30; Video quality: 100% 8bit RGB; Video format: uncompressed AVI; Camera distance: 20 cm; Illumination: standard light bulb XX cd/m2

PublicationCorrelation (F/W/C)*|HR rPPG minus HR reference| Face (R/L/M)**|HR rPPG minus HR reference| Wrist (R/L/M)**|HR rPPG minus HR reference| Calf (R/L/M)**AUC Face (R-L/R-M/L-M)AUC Wrist (R-L/R-M/L-M)AUC Calf (R-L/R-M/L-M)Download sofware
van der Kooij & Naber, 20190.97/0.50/0.272.34/5.91/5.6029.75/16.11/18.7836.83/19.47/29.070.92/1.00/0.890.66/0.87/0.790.53/0.57/0.53
* F = face, W = wrist, C = calf
** HR = heart rate;
R = during rest; L = after light exercise; M = after moderate exercise



Here comes the next table for rPPG benchmarks on additional video databases
PublicationCorrelationetc.
Surname, yearXetc.