Video-based Heart Rate Estimation with Angle-guided ROI Optimization and Graph Signal Denoising

Boqiu Shen; Gan Pei; Junhao Ning; Menghan Hu; Yan Zhu

arxiv: 2604.11395 · v1 · submitted 2026-04-13 · 💻 cs.CV

Video-based Heart Rate Estimation with Angle-guided ROI Optimization and Graph Signal Denoising

Gan Pei , Junhao Ning , Boqiu Shen , Yan Zhu , Menghan Hu This is my paper

Pith reviewed 2026-05-10 15:43 UTC · model grok-4.3

classification 💻 cs.CV

keywords rPPGremote photoplethysmographyheart rate estimationmotion artifactsgraph signal processingROI optimizationfacial video analysisdenoising

0 comments

The pith

Two plug-and-play modules using ROI-camera angles and graph signal processing reduce errors in video-based heart rate measurement during motion.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper introduces two modules to make remote photoplethysmography more robust to facial motions like speaking and head shaking. One module quantifies the angles between selected facial regions and the camera to refine affected signals and capture overall movement. The other module treats multiple regions as a graph and applies joint signal processing to remove motion noise while keeping physiological information. When added to existing reflection-based rPPG methods, the pair lowers average mean absolute error by 20.38 percent across three public datasets. A reader would care because reliable non-contact heart rate tracking becomes feasible in everyday settings without requiring subjects to stay still.

Core claim

The central claim is that the Angle-guided ROI Adaptive Optimization module, which quantifies ROI-Camera angles to refine motion-affected signals and capture global motion, combined with the Multi-region Joint Graph Signal Denoising module, which jointly models intra- and inter-regional ROI signals using graph signal processing to suppress motion artifacts, produces markedly lower mean absolute error when integrated with reflection model-based rPPG methods.

What carries the argument

The two plug-and-play modules: angle-guided ROI adaptive optimization that quantifies camera angles to adjust signals, and multi-region joint graph signal denoising that models relationships across facial regions to filter artifacts.

If this is right

The modules integrate directly with existing reflection model-based rPPG methods without retraining.
Performance gains hold across three public datasets containing speaking and head-shaking motions.
Ablation tests show each module contributes independently to the overall error reduction.
The approach supports more practical deployment of video heart rate estimation in motion-heavy environments.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The same angle and graph techniques could be tested on other remote vital-sign tasks such as respiration rate from video.
Extending the graph to include temporal edges across video frames might further stabilize estimates during prolonged motion.
The method may lower the barrier to using rPPG in telemedicine or fitness applications where subjects move naturally.

Load-bearing premise

That quantifying ROI-Camera angles accurately captures and mitigates motion artifacts without introducing new biases, and that the graph denoising preserves true physiological signals across regions.

What would settle it

Apply the modules to a dataset with simultaneous ECG ground truth and known extreme head motions, then disable the angle quantification or graph edges and check whether the 20 percent MAE reduction disappears or reverses.

read the original abstract

Remote photoplethysmography (rPPG) enables non-contact heart rate measurement from facial videos, but its performance is significantly degraded by facial motions such as speaking and head shaking. To address this issue, we propose two plug-and-play modules. The Angle-guided ROI Adaptive Optimization module quantifies ROI-Camera angles to refine motion-affected signals and capture global motion, while the Multi-region Joint Graph Signal Denoising module jointly models intra- and inter-regional ROI signals using graph signal processing to suppress motion artifacts. The modules are compatible with reflection model-based rPPG methods and validated on three public datasets. Results show that jointly use markedly reduces MAE, with an average decrease of 20.38\% over the baseline, while ablation studies confirm the effectiveness of each module. The work demonstrates the potential of angle-guided optimization and graph-based denoising to enhance rPPG performance in motion scenarios.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The paper adds angle-guided ROI optimization and joint graph denoising to rPPG pipelines and reports a 20% average MAE drop on motion-heavy tests across three datasets.

read the letter

The main takeaway is that these two plug-and-play modules together cut mean absolute error by about 20 percent on average versus baseline when facial motion is present. The angle module refines regions of interest by measuring ROI-camera angles to handle global motion, while the graph module jointly denoises signals across multiple regions using graph signal processing. Both are meant to slot into existing reflection-model rPPG methods without major changes. The paper backs this with results on three public datasets and ablations that isolate each module's contribution, plus tests on speaking and head-shaking cases. That combination and the empirical numbers are the clearest new pieces. The work does well by staying practical and showing concrete gains rather than just proposing ideas in theory. The ablations give some reassurance that the improvements are not just from one part carrying the other. On the soft spots, the angle quantification rests on the idea that those measurements capture motion artifacts cleanly without adding their own bias, and the graph step preserves the underlying heart-rate signal. The ablations test this to some degree, but more detail on graph construction, parameter sensitivity, or comparisons to simpler filters would make the case tighter. The reported average improvement is useful, yet without error bars or per-dataset breakdowns it is harder to judge how consistent the gains are across subjects or conditions. This paper is for people working on non-contact vital-signs monitoring who need better motion handling in real settings. Readers in computer vision or biomedical signal processing who build or adapt rPPG systems would get direct value from the modules and the dataset results. It has enough empirical grounding and clear methods to deserve a serious referee. I would recommend sending it to peer review. The core claim is testable and the experiments are on standard data, so referees can check the details and push for tighter comparisons where needed.

Referee Report

1 major / 2 minor

Summary. The paper proposes two plug-and-play modules to improve remote photoplethysmography (rPPG) heart rate estimation from facial videos under motion artifacts: an Angle-guided ROI Adaptive Optimization module that quantifies ROI-camera angles to refine motion-affected signals and capture global motion, and a Multi-region Joint Graph Signal Denoising module that jointly models intra- and inter-regional ROI signals via graph signal processing to suppress artifacts. These modules are presented as compatible with existing reflection model-based rPPG methods and are evaluated on three public datasets, with the joint use claimed to reduce MAE by an average of 20.38% over baseline, supported by ablation studies confirming each module's contribution.

Significance. If the empirical MAE reductions are reproducible with full experimental details, the work offers a practical, modular approach to mitigating motion artifacts in non-contact HR estimation, potentially improving robustness in real-world scenarios like speaking or head shaking. The angle-guided and graph-based techniques provide a distinct perspective on ROI optimization and signal denoising, with the plug-and-play design facilitating adoption; however, the current lack of verification details limits assessment of broader impact.

major comments (1)

The abstract and experimental sections report a 20.38% average MAE reduction and ablation results but omit key details on data splits, error bars, statistical significance tests, and exact hyperparameter choices or post-hoc selections; these omissions directly limit verification of the central empirical claim and the weakest assumption that angle quantification and graph denoising mitigate artifacts without introducing bias.

minor comments (2)

Notation for ROI angles and graph construction in the method description could be clarified with explicit equations or pseudocode to improve reproducibility.
Figure captions and table legends should explicitly state the baseline methods and dataset motion conditions for easier cross-reference with the reported improvements.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the constructive review and positive recommendation for minor revision. We agree that additional details are needed to support reproducibility and verification of our empirical claims, and we will incorporate them in the revised manuscript.

read point-by-point responses

Referee: The abstract and experimental sections report a 20.38% average MAE reduction and ablation results but omit key details on data splits, error bars, statistical significance tests, and exact hyperparameter choices or post-hoc selections; these omissions directly limit verification of the central empirical claim and the weakest assumption that angle quantification and graph denoising mitigate artifacts without introducing bias.

Authors: We appreciate this feedback and fully acknowledge that the current version lacks sufficient details for independent verification. In the revised manuscript, we will expand the experimental section (and update the abstract if space allows) to include: (1) explicit descriptions of the data splits used for each of the three public datasets, including whether they are subject-independent; (2) error bars (mean ± standard deviation) for all MAE results, computed across subjects or multiple runs; (3) statistical significance tests such as paired t-tests or Wilcoxon signed-rank tests with p-values comparing our full method and ablations against baselines; and (4) a dedicated table listing all hyperparameters with exact values, ranges, and the selection procedure (including any post-hoc choices). Regarding the assumption of no introduced bias, the Angle-guided ROI Adaptive Optimization module uses geometric angle quantification solely to select and weight ROIs for better motion capture, without altering signal amplitudes or introducing systematic shifts. The Multi-region Joint Graph Signal Denoising module applies graph-based low-pass filtering to suppress high-frequency motion components while preserving the low-frequency physiological signal, as validated by our frequency-domain analysis. Ablation results already show gains in motion scenarios without degradation on low-motion subsets, supporting that bias is not introduced. We will add an explicit discussion subsection on bias analysis and mitigation in the revision. revision: yes

Circularity Check

0 steps flagged

No significant circularity detected

full rationale

The paper proposes two empirical plug-and-play modules (angle-guided ROI optimization and multi-region graph signal denoising) as compatible additions to existing reflection-model rPPG methods. Its central claims rest on experimental MAE reductions (20.38% average) and ablations across three public datasets under motion conditions, not on any mathematical derivation, first-principles prediction, or self-referential definition. No equations or steps reduce by construction to fitted inputs, self-citations, or renamed known results; the argument structure is self-contained via external dataset validation and does not invoke load-bearing uniqueness theorems or ansatzes from prior author work.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract-only review provides no identifiable free parameters, axioms, or invented entities; modules are framed as plug-and-play extensions without new postulated entities.

pith-pipeline@v0.9.0 · 5457 in / 943 out tokens · 23595 ms · 2026-05-10T15:43:21.408226+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

22 extracted references · 22 canonical work pages · 1 internal anchor

[1]

Video-based Heart Rate Estimation with Angle-guided ROI Optimization and Graph Signal Denoising

INTRODUCTION Remote photoplethysmography (rPPG) is a non-contact, non- invasive, and cost-efficient optical technique that derives blood volume pulse (BVP) signals from skin regions in RGB video [1]. It enables the estimation of vital signs such as heart rate (HR) and respiration, with broad applicability in telemedicine, continuous health monitoring, and...

work page internal anchor Pith review Pith/arXiv arXiv 2026
[2]

Angle-guided ROI Adaptive Optimization The Pipeline of the proposed method is illustrated in Fig

METHOD 2.1. Angle-guided ROI Adaptive Optimization The Pipeline of the proposed method is illustrated in Fig. 2. MediaPipe provides dense 3D coordinates of 468 facial land- marks. Leveraging facial midline symmetry, 60 landmarks are selected on the forehead, cheeks, and chin; for each selected landmark, a20×20-pixel rectangular ROI is con- structed center...

work page
[3]

Performance comparison of baseline rPPG methods with angle-guided ROI module (AGROI) and multi-region joint graph signal denoising module (GSD) on three pub- lic datasets

EXPERIMENT AND RESULTS Since the modules essentially target the pre- and post- processing stages of HR estimation, their effectiveness is Table 1. Performance comparison of baseline rPPG methods with angle-guided ROI module (AGROI) and multi-region joint graph signal denoising module (GSD) on three pub- lic datasets. S-E: Signal Extraction; S-P: Signal Pr...

work page
[4]

CONCLUSION In this work, we propose two plug-and-play modules that an Angle-guided ROI Adaptive Optimization module and a Multi-region Joint Graph Signal Denoising module to en- hance video-based HR estimation. The first module adap- tively refines ROI signals by leveraging head pose variations while preserving signal redundancy, and the second exploits i...

work page
[5]

Remote plethysmographic imaging using ambi- ent light.,

Wim Verkruysse, Lars O Svaasand, and J Stuart Nel- son, “Remote plethysmographic imaging using ambi- ent light.,”Optics Express, vol. 16, no. 26, pp. 21434– 21445, 2008

work page 2008
[6]

Graph-based denoising for respiration and heart rate estimation during sleep in thermal video,

Cheng Yang, Menghan Hu, Guangtao Zhai, and Xiao- Ping Zhang, “Graph-based denoising for respiration and heart rate estimation during sleep in thermal video,” IEEE Internet of Things Journal, vol. 9, no. 17, pp. 15697–15713, 2022

work page 2022
[7]

Iden- tification of deep breath while moving forward based on multiple body regions and graph signal analysis,

Yunlu Wang, Cheng Yang, Menghan Hu, Jian Zhang, Qingli Li, Guangtao Zhai, and Xiao-Ping Zhang, “Iden- tification of deep breath while moving forward based on multiple body regions and graph signal analysis,” inICASSP 2021-2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2021, pp. 7958–7962

work page 2021
[8]

A survey of remote optical photoplethysmographic imaging methods,

Daniel J McDuff, Justin R Estepp, Alyssa M Piasecki, and Ethan B Blackford, “A survey of remote optical photoplethysmographic imaging methods,” in2015 37th Annual International Conference of the IEEE Engineer- ing in Medicine and Biology Society (EMBC), 2015, pp. 6398–6404

work page 2015
[9]

Improved mo- tion robustness of remote-ppg by using the blood vol- ume pulse signature,

Gerard De Haan and Arno Van Leest, “Improved mo- tion robustness of remote-ppg by using the blood vol- ume pulse signature,”Physiological Measurement, vol. 35, no. 9, pp. 1913, 2014

work page 1913
[10]

Algorithmic principles of remote ppg,

Wenjin Wang, Albertus C Den Brinker, Sander Stuijk, and Gerard De Haan, “Algorithmic principles of remote ppg,”IEEE Transactions on Biomedical Engineering, vol. 64, no. 7, pp. 1479–1491, 2016

work page 2016
[11]

Face2ppg: An unsupervised pipeline for blood volume pulse extraction from faces,

Constantino Alvarez Casado and Miguel Bordallo L´opez, “Face2ppg: An unsupervised pipeline for blood volume pulse extraction from faces,”IEEE Journal of Biomedical and Health Informatics, vol. 27, no. 11, pp. 5530–5541, 2023

work page 2023
[12]

Efficientphys: Enabling simple, fast and accurate camera-based cardiac measurement,

Xin Liu, Brian Hill, Ziheng Jiang, Shwetak Patel, and Daniel McDuff, “Efficientphys: Enabling simple, fast and accurate camera-based cardiac measurement,” in Proceedings of the IEEE/CVF Conference on Com- puter Vision and Pattern Recognition (CVPR), 2023, pp. 5008–5017

work page 2023
[13]

Rhythm- former: Extracting patterned rppg signals based on pe- riodic sparse attention,

Bochao Zou, Zizheng Guo, Jiansheng Chen, Junbao Zhuo, Weiran Huang, and Huimin Ma, “Rhythm- former: Extracting patterned rppg signals based on pe- riodic sparse attention,”Pattern Recognition, vol. 164, pp. 111511, 2025

work page 2025
[14]

Phys- former: Facial video-based physiological measurement with temporal difference transformer,

Zitong Yu, Yuming Shen, Jingang Shi, Hengshuang Zhao, Philip HS Torr, and Guoying Zhao, “Phys- former: Facial video-based physiological measurement with temporal difference transformer,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2022, pp. 4186–4196

work page 2022
[15]

Video-based heart rate estimation with spectrogram signal quality ranking and fusion,

Rencheng Song, Zhenzhou Du, Juan Cheng, Chang Li, and Xuezhi Yang, “Video-based heart rate estimation with spectrogram signal quality ranking and fusion,” Biomedical Signal Processing and Control, vol. 100, pp. 107094, 2025

work page 2025
[16]

Orientation-conditioned facial texture mapping for video-based facial remote photoplethysmography esti- mation,

Sam Cantrill, David Ahmedt-Aristizabal, Lars Peters- son, Hanna Suominen, and Mohammad Ali Armin, “Orientation-conditioned facial texture mapping for video-based facial remote photoplethysmography esti- mation,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2024, pp. 354–363

work page 2024
[17]

Ex- ploiting spatial redundancy of image sensor for motion robust rppg,

Wenjin Wang, Sander Stuijk, and Gerard De Haan, “Ex- ploiting spatial redundancy of image sensor for motion robust rppg,”IEEE Transactions on Biomedical Engi- neering, vol. 62, no. 2, pp. 415–425, 2014

work page 2014
[18]

Estimating heart rate and rhythm via 3d motion track- ing in depth video,

Cheng Yang, Gene Cheung, and Vladimir Stankovic, “Estimating heart rate and rhythm via 3d motion track- ing in depth video,”IEEE Transactions on Multimedia, vol. 19, no. 7, pp. 1625–1636, 2017

work page 2017
[19]

Non-contact video-based pulse rate measure- ment on a mobile service robot,

Ronny Stricker, Steffen M ¨uller, and Horst-Michael Gross, “Non-contact video-based pulse rate measure- ment on a mobile service robot,” inThe 23rd IEEE In- ternational Symposium on Robot and Human Interactive Communication, 2014, pp. 1056–1062

work page 2014
[20]

A reproducible study on remote heart rate mea- surement,

Guillaume Heusch, Andr ´e Anjos, and S ´ebastien Mar- cel, “A reproducible study on remote heart rate mea- surement,”arXiv, 2017

work page 2017
[21]

Mmpd: Multi-domain mobile video physiology dataset,

Jiankai Tang, Kequan Chen, Yuntao Wang, Yuanchun Shi, Shwetak Patel, Daniel McDuff, and Xin Liu, “Mmpd: Multi-domain mobile video physiology dataset,” in2023 45th Annual International Conference of the IEEE Engineering in Medicine & Biology Society (EMBC), 2023, pp. 1–5

work page 2023
[22]

rppg-toolbox: Deep remote ppg toolbox,

Xin Liu, Girish Narayanswamy, Akshay Paruchuri, Xi- aoyu Zhang, Jiankai Tang, Yuzhe Zhang, Roni Sen- gupta, Shwetak Patel, Yuntao Wang, and Daniel Mc- Duff, “rppg-toolbox: Deep remote ppg toolbox,”Ad- vances in Neural Information Processing Systems, vol. 36, pp. 68485–68510, 2023

work page 2023

[1] [1]

Video-based Heart Rate Estimation with Angle-guided ROI Optimization and Graph Signal Denoising

INTRODUCTION Remote photoplethysmography (rPPG) is a non-contact, non- invasive, and cost-efficient optical technique that derives blood volume pulse (BVP) signals from skin regions in RGB video [1]. It enables the estimation of vital signs such as heart rate (HR) and respiration, with broad applicability in telemedicine, continuous health monitoring, and...

work page internal anchor Pith review Pith/arXiv arXiv 2026

[2] [2]

Angle-guided ROI Adaptive Optimization The Pipeline of the proposed method is illustrated in Fig

METHOD 2.1. Angle-guided ROI Adaptive Optimization The Pipeline of the proposed method is illustrated in Fig. 2. MediaPipe provides dense 3D coordinates of 468 facial land- marks. Leveraging facial midline symmetry, 60 landmarks are selected on the forehead, cheeks, and chin; for each selected landmark, a20×20-pixel rectangular ROI is con- structed center...

work page

[3] [3]

Performance comparison of baseline rPPG methods with angle-guided ROI module (AGROI) and multi-region joint graph signal denoising module (GSD) on three pub- lic datasets

EXPERIMENT AND RESULTS Since the modules essentially target the pre- and post- processing stages of HR estimation, their effectiveness is Table 1. Performance comparison of baseline rPPG methods with angle-guided ROI module (AGROI) and multi-region joint graph signal denoising module (GSD) on three pub- lic datasets. S-E: Signal Extraction; S-P: Signal Pr...

work page

[4] [4]

CONCLUSION In this work, we propose two plug-and-play modules that an Angle-guided ROI Adaptive Optimization module and a Multi-region Joint Graph Signal Denoising module to en- hance video-based HR estimation. The first module adap- tively refines ROI signals by leveraging head pose variations while preserving signal redundancy, and the second exploits i...

work page

[5] [5]

Remote plethysmographic imaging using ambi- ent light.,

Wim Verkruysse, Lars O Svaasand, and J Stuart Nel- son, “Remote plethysmographic imaging using ambi- ent light.,”Optics Express, vol. 16, no. 26, pp. 21434– 21445, 2008

work page 2008

[6] [6]

Graph-based denoising for respiration and heart rate estimation during sleep in thermal video,

Cheng Yang, Menghan Hu, Guangtao Zhai, and Xiao- Ping Zhang, “Graph-based denoising for respiration and heart rate estimation during sleep in thermal video,” IEEE Internet of Things Journal, vol. 9, no. 17, pp. 15697–15713, 2022

work page 2022

[7] [7]

Iden- tification of deep breath while moving forward based on multiple body regions and graph signal analysis,

Yunlu Wang, Cheng Yang, Menghan Hu, Jian Zhang, Qingli Li, Guangtao Zhai, and Xiao-Ping Zhang, “Iden- tification of deep breath while moving forward based on multiple body regions and graph signal analysis,” inICASSP 2021-2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2021, pp. 7958–7962

work page 2021

[8] [8]

A survey of remote optical photoplethysmographic imaging methods,

Daniel J McDuff, Justin R Estepp, Alyssa M Piasecki, and Ethan B Blackford, “A survey of remote optical photoplethysmographic imaging methods,” in2015 37th Annual International Conference of the IEEE Engineer- ing in Medicine and Biology Society (EMBC), 2015, pp. 6398–6404

work page 2015

[9] [9]

Improved mo- tion robustness of remote-ppg by using the blood vol- ume pulse signature,

Gerard De Haan and Arno Van Leest, “Improved mo- tion robustness of remote-ppg by using the blood vol- ume pulse signature,”Physiological Measurement, vol. 35, no. 9, pp. 1913, 2014

work page 1913

[10] [10]

Algorithmic principles of remote ppg,

Wenjin Wang, Albertus C Den Brinker, Sander Stuijk, and Gerard De Haan, “Algorithmic principles of remote ppg,”IEEE Transactions on Biomedical Engineering, vol. 64, no. 7, pp. 1479–1491, 2016

work page 2016

[11] [11]

Face2ppg: An unsupervised pipeline for blood volume pulse extraction from faces,

Constantino Alvarez Casado and Miguel Bordallo L´opez, “Face2ppg: An unsupervised pipeline for blood volume pulse extraction from faces,”IEEE Journal of Biomedical and Health Informatics, vol. 27, no. 11, pp. 5530–5541, 2023

work page 2023

[12] [12]

Efficientphys: Enabling simple, fast and accurate camera-based cardiac measurement,

Xin Liu, Brian Hill, Ziheng Jiang, Shwetak Patel, and Daniel McDuff, “Efficientphys: Enabling simple, fast and accurate camera-based cardiac measurement,” in Proceedings of the IEEE/CVF Conference on Com- puter Vision and Pattern Recognition (CVPR), 2023, pp. 5008–5017

work page 2023

[13] [13]

Rhythm- former: Extracting patterned rppg signals based on pe- riodic sparse attention,

Bochao Zou, Zizheng Guo, Jiansheng Chen, Junbao Zhuo, Weiran Huang, and Huimin Ma, “Rhythm- former: Extracting patterned rppg signals based on pe- riodic sparse attention,”Pattern Recognition, vol. 164, pp. 111511, 2025

work page 2025

[14] [14]

Phys- former: Facial video-based physiological measurement with temporal difference transformer,

Zitong Yu, Yuming Shen, Jingang Shi, Hengshuang Zhao, Philip HS Torr, and Guoying Zhao, “Phys- former: Facial video-based physiological measurement with temporal difference transformer,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2022, pp. 4186–4196

work page 2022

[15] [15]

Video-based heart rate estimation with spectrogram signal quality ranking and fusion,

Rencheng Song, Zhenzhou Du, Juan Cheng, Chang Li, and Xuezhi Yang, “Video-based heart rate estimation with spectrogram signal quality ranking and fusion,” Biomedical Signal Processing and Control, vol. 100, pp. 107094, 2025

work page 2025

[16] [16]

Orientation-conditioned facial texture mapping for video-based facial remote photoplethysmography esti- mation,

Sam Cantrill, David Ahmedt-Aristizabal, Lars Peters- son, Hanna Suominen, and Mohammad Ali Armin, “Orientation-conditioned facial texture mapping for video-based facial remote photoplethysmography esti- mation,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2024, pp. 354–363

work page 2024

[17] [17]

Ex- ploiting spatial redundancy of image sensor for motion robust rppg,

Wenjin Wang, Sander Stuijk, and Gerard De Haan, “Ex- ploiting spatial redundancy of image sensor for motion robust rppg,”IEEE Transactions on Biomedical Engi- neering, vol. 62, no. 2, pp. 415–425, 2014

work page 2014

[18] [18]

Estimating heart rate and rhythm via 3d motion track- ing in depth video,

Cheng Yang, Gene Cheung, and Vladimir Stankovic, “Estimating heart rate and rhythm via 3d motion track- ing in depth video,”IEEE Transactions on Multimedia, vol. 19, no. 7, pp. 1625–1636, 2017

work page 2017

[19] [19]

Non-contact video-based pulse rate measure- ment on a mobile service robot,

Ronny Stricker, Steffen M ¨uller, and Horst-Michael Gross, “Non-contact video-based pulse rate measure- ment on a mobile service robot,” inThe 23rd IEEE In- ternational Symposium on Robot and Human Interactive Communication, 2014, pp. 1056–1062

work page 2014

[20] [20]

A reproducible study on remote heart rate mea- surement,

Guillaume Heusch, Andr ´e Anjos, and S ´ebastien Mar- cel, “A reproducible study on remote heart rate mea- surement,”arXiv, 2017

work page 2017

[21] [21]

Mmpd: Multi-domain mobile video physiology dataset,

Jiankai Tang, Kequan Chen, Yuntao Wang, Yuanchun Shi, Shwetak Patel, Daniel McDuff, and Xin Liu, “Mmpd: Multi-domain mobile video physiology dataset,” in2023 45th Annual International Conference of the IEEE Engineering in Medicine & Biology Society (EMBC), 2023, pp. 1–5

work page 2023

[22] [22]

rppg-toolbox: Deep remote ppg toolbox,

Xin Liu, Girish Narayanswamy, Akshay Paruchuri, Xi- aoyu Zhang, Jiankai Tang, Yuzhe Zhang, Roni Sen- gupta, Shwetak Patel, Yuntao Wang, and Daniel Mc- Duff, “rppg-toolbox: Deep remote ppg toolbox,”Ad- vances in Neural Information Processing Systems, vol. 36, pp. 68485–68510, 2023

work page 2023