Reliability factor for accurate remote PPG systems

The significance of contactless vital data monitoring has gained increasing attention in recent years, with light playing a crucial role in its implementation. Visible light, capable of penetrating thin body regions, enables inferences to be drawn regarding changes in volume, blood color, and pulse rates based on alterations in light intensity. Additionally, vital signs can be ascertained from light reflected off the skin, which can be captured by a standard camera when utilizing ambient light. Remote photoplethysmography (rPPG), a non-contact method that functions effectively over greater distances, remains highly susceptible to lighting conditions, individual physical attributes, and motion artifacts. Consequently, rPPG signal quality ranges from highly accurate to extremely poor, resulting in an unstable, non-reproducible, and challenging-to-implement technology. In this conference paper, we provide a comprehensive overview of the latest rPPG advancements, introduce annotated video databases for quality validation, and propose a reliability factor to address the issue of fluctuating signal quality. We demonstrate the potential of our approach through the CareCam application, a workplaceoriented solution employing a standard webcam to acquire vital data in real-time.