Assessing the Feasibility of Remote Photoplethysmography Through Videocalls: A Study of Network and Computing Constraints
Álvarez Casado, Constantino; Nguyen, Le; Silvén, Olli; Bordallo López, Miguel (2023-04-27)
Springer
27.04.2023
Álvarez Casado, C., Nguyen, L., Silvén, O., Bordallo López, M. (2023). Assessing the Feasibility of Remote Photoplethysmography Through Videocalls: A Study of Network and Computing Constraints. In: Gade, R., Felsberg, M., Kämäräinen, JK. (eds) Image Analysis. SCIA 2023. Lecture Notes in Computer Science, vol 13886. Springer, Cham. https://doi.org/10.1007/978-3-031-31438-4_38
https://rightsstatements.org/vocab/InC/1.0/
© 2023 The Author(s), under exclusive license to Springer Nature Switzerland AG
https://rightsstatements.org/vocab/InC/1.0/
© 2023 The Author(s), under exclusive license to Springer Nature Switzerland AG
https://rightsstatements.org/vocab/InC/1.0/
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:oulu-202404182838
https://urn.fi/URN:NBN:fi:oulu-202404182838
Tiivistelmä
Abstract
Remote photoplethysmography (rPPG) is a promising non-invasive technique for measuring vital signs remotely, such as through videocalls. However, network and computing constraints can significantly compromise its accuracy. In this study, we evaluated the effects of these constraints on rPPG methods using four reference datasets and a standard unsupervised rPPG signal extraction pipeline. Our experiments simulated the impact of frame dropping, streaming video at different resolutions and frame rates, and other resource limitations. We found that these constraints can significantly degrade rPPG accuracy, but implementing specific strategies (such as reconstructing the signal in the receiver) can mitigate these effects. For example, with a 20% of frame loss, our proposed strategies reduced the MAE increase from 539% to 29%. These findings highlight the importance of considering network and computing constraints in rPPG applications.
Remote photoplethysmography (rPPG) is a promising non-invasive technique for measuring vital signs remotely, such as through videocalls. However, network and computing constraints can significantly compromise its accuracy. In this study, we evaluated the effects of these constraints on rPPG methods using four reference datasets and a standard unsupervised rPPG signal extraction pipeline. Our experiments simulated the impact of frame dropping, streaming video at different resolutions and frame rates, and other resource limitations. We found that these constraints can significantly degrade rPPG accuracy, but implementing specific strategies (such as reconstructing the signal in the receiver) can mitigate these effects. For example, with a 20% of frame loss, our proposed strategies reduced the MAE increase from 539% to 29%. These findings highlight the importance of considering network and computing constraints in rPPG applications.
Kokoelmat
- Avoin saatavuus [37606]