VitalCamSet - a dataset for Photoplethysmography Imaging

In recent years, several approaches for Photoplethysmography Imaging have been proposed. However, the so far proposed approaches have been evaluated either on non-public datasets or on such proposed for different research scope. Therefore, we present VitalCamSet, a new dataset for PPGI allowing comparative evaluation of different algorithmic approaches. The VitalCamSet contains videos of 26 subjects with 10 different scenarios and a length of two minutes each, recorded simultaneously by a color camera and a monochrome near-infrared camera. The video data was synchronized using the cross correlation of LED signals, captured in the videos, as well as in the light sensor of the vital sign reference system. In addition, signal quality of the videos was analyzed using a SNR (Signal-to-Noise-Ratio) metric. The dataset is made publicly available and we encourage PPGI researchers to use it for testing their developed PPGI approaches.