Deception Detection In virtual sales meetings

The aim of this project is to determine vital signs (heart rate, respiratory rate & heart rate variability) on the basis of contactless camera-based measurement of PPG signals in order to enable the derivation of clinically relevant indicators. The starting point is our preliminary work on camera-based recording of heart rate and respiratory rate. However, the determination of heart rate variability requires a more interference-free analysis of the PPG signal than the current state of the art allows. This requires the exact temporal localization of the heartbeats. In order to achieve this, new methods are to be researched in this project with which the vital signs can be robustly derived from the video PPG signals. The newly developed machine learning approaches for image and signal processing will be investigated both on the RGB videos commonly used in the state of the art and on various multispectral bands in the visible and near-infrared light range. The aim is to specifically learn a motion and illumination artifact-invariant measurement of vital signs.