Improving Video-based Heart Rate and Respiratory Rate Estimation via Pulse-Respiration Quotient

Yuzhuo Ren, Braeden Syrnyk, Niranjan Avadhanam
Proceedings of the 1st Workshop on Healthcare AI and COVID-19, ICML 2022, PMLR 184:136-145, 2022.

Abstract

Remote physiological measurement, \textit{e.g.}, heart rate and respiratory rate measurement, becomes more and more important when contact instrument-based measurement is inaccessible and non-preferable under the COVID-19 pandemic. Non-contact camera based physiological measurement enables Telehealth, remote health monitoring and smart hospital applications. Remote physiological signal measurement has challenges such as environment illumination variations, head motion, facial expression, etc. We propose a convolutional neural network to jointly estimate heart rate and respiratory rate with camera video as input in a multitask fashion, which leverages the correlation between heart rate and respiratory rate. Specifically, we propose a novel loss function which integrates the frequency correlation between heart rate and respiratory rate to improve robustness of both heart rate and respiratory rate estimation. Furthermore, we propose a post processing filter based on correlation between heart rate and respiratory rate which further improve prediction accuracy. Extensive experiments demonstrate that our proposed system significantly improves heart rate and respiratory rate measurement accuracy.

Cite this Paper

BibTeX
@InProceedings{pmlr-v184-ren22a, title = {Improving Video-based Heart Rate and Respiratory Rate Estimation via Pulse-Respiration Quotient}, author = {Ren, Yuzhuo and Syrnyk, Braeden and Avadhanam, Niranjan}, booktitle = {Proceedings of the 1st Workshop on Healthcare AI and COVID-19, ICML 2022}, pages = {136--145}, year = {2022}, editor = {Xu, Peng and Zhu, Tingting and Zhu, Pengkai and Clifton, David A. and Belgrave, Danielle and Zhang, Yuanting}, volume = {184}, series = {Proceedings of Machine Learning Research}, month = {22 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v184/ren22a/ren22a.pdf}, url = {https://proceedings.mlr.press/v184/ren22a.html}, abstract = {Remote physiological measurement, \textit{e.g.}, heart rate and respiratory rate measurement, becomes more and more important when contact instrument-based measurement is inaccessible and non-preferable under the COVID-19 pandemic. Non-contact camera based physiological measurement enables Telehealth, remote health monitoring and smart hospital applications. Remote physiological signal measurement has challenges such as environment illumination variations, head motion, facial expression, etc. We propose a convolutional neural network to jointly estimate heart rate and respiratory rate with camera video as input in a multitask fashion, which leverages the correlation between heart rate and respiratory rate. Specifically, we propose a novel loss function which integrates the frequency correlation between heart rate and respiratory rate to improve robustness of both heart rate and respiratory rate estimation. Furthermore, we propose a post processing filter based on correlation between heart rate and respiratory rate which further improve prediction accuracy. Extensive experiments demonstrate that our proposed system significantly improves heart rate and respiratory rate measurement accuracy.} }
Endnote
%0 Conference Paper %T Improving Video-based Heart Rate and Respiratory Rate Estimation via Pulse-Respiration Quotient %A Yuzhuo Ren %A Braeden Syrnyk %A Niranjan Avadhanam %B Proceedings of the 1st Workshop on Healthcare AI and COVID-19, ICML 2022 %C Proceedings of Machine Learning Research %D 2022 %E Peng Xu %E Tingting Zhu %E Pengkai Zhu %E David A. Clifton %E Danielle Belgrave %E Yuanting Zhang %F pmlr-v184-ren22a %I PMLR %P 136--145 %U https://proceedings.mlr.press/v184/ren22a.html %V 184 %X Remote physiological measurement, \textit{e.g.}, heart rate and respiratory rate measurement, becomes more and more important when contact instrument-based measurement is inaccessible and non-preferable under the COVID-19 pandemic. Non-contact camera based physiological measurement enables Telehealth, remote health monitoring and smart hospital applications. Remote physiological signal measurement has challenges such as environment illumination variations, head motion, facial expression, etc. We propose a convolutional neural network to jointly estimate heart rate and respiratory rate with camera video as input in a multitask fashion, which leverages the correlation between heart rate and respiratory rate. Specifically, we propose a novel loss function which integrates the frequency correlation between heart rate and respiratory rate to improve robustness of both heart rate and respiratory rate estimation. Furthermore, we propose a post processing filter based on correlation between heart rate and respiratory rate which further improve prediction accuracy. Extensive experiments demonstrate that our proposed system significantly improves heart rate and respiratory rate measurement accuracy.
APA
Ren, Y., Syrnyk, B. & Avadhanam, N.. (2022). Improving Video-based Heart Rate and Respiratory Rate Estimation via Pulse-Respiration Quotient. Proceedings of the 1st Workshop on Healthcare AI and COVID-19, ICML 2022, in Proceedings of Machine Learning Research 184:136-145 Available from https://proceedings.mlr.press/v184/ren22a.html.

Related Material