Optimize-at-Capture: Highly-adaptive Exposure Controlling for In-Vehicle Non-contact Heart-rate Monitoring

Caifeng Shan; Jieying Wang; Wenjin Wang; Xinqi Cai

arxiv: 2605.04397 · v1 · submitted 2026-05-06 · 💻 cs.CV · cs.SY· eess.SY

Optimize-at-Capture: Highly-adaptive Exposure Controlling for In-Vehicle Non-contact Heart-rate Monitoring

Jieying Wang , Xinqi Cai , Caifeng Shan , Wenjin Wang This is my paper

Pith reviewed 2026-05-08 18:07 UTC · model grok-4.3

classification 💻 cs.CV cs.SYeess.SY

keywords remote photoplethysmographyexposure controlin-vehicle monitoringheart rate estimationadaptive imagingdriving conditionsphysiological signal

0 comments

The pith

Predictive modeling of skin reflections lets in-vehicle cameras hold faces in the optimal brightness range for remote heart-rate tracking.

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

The paper demonstrates that fixed exposure and standard auto-exposure fail to keep facial skin at steady brightness when light shifts quickly during driving, which destroys the tiny color pulses used for remote heart-rate measurement. The authors introduce a method that uses past skin reflection data to forecast the right exposure settings for the next frame, keeping the region of interest inside the camera's best dynamic range for signal quality. Tests on real driving footage show this cuts average heart-rate error from 14.1 to 7.79 beats per minute and raises the success rate from 25 to 57 percent.

Core claim

By proactively adjusting exposure parameters based on predictive modeling of historical skin reflections and optimizing specifically for rPPG measurement, the skin region of interest remains within the optimal dynamic range, enabling more accurate non-contact heart-rate monitoring under the rapidly changing illumination conditions typical of vehicle interiors.

What carries the argument

The highly-adaptive exposure controlling framework that predicts future lighting from historical skin reflections to set camera exposure parameters in real time.

If this is right

It reduces mean absolute error in heart rate estimates by 6.31 bpm compared to standard methods.
It increases the success rate of valid measurements by 32.3 percentage points in challenging driving conditions.
It shows clear gains in both low-light rainy and high-glare sunny scenarios.
The ExpDrive dataset provides synchronized facial video and ECG from 48 subjects for further development of in-vehicle monitoring.

Where Pith is reading between the lines

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

The same predictive approach might help other camera-based vital sign measurements that depend on stable pixel values in changing light.
The ExpDrive dataset could serve as a benchmark for testing exposure strategies combined with motion compensation or multi-camera setups.
Real-time vehicle implementation would require balancing prediction accuracy against available processing time per frame.

Load-bearing premise

The system assumes that patterns in past skin reflections can predict the immediate next lighting condition accurately enough to set exposure without adding lag or artifacts that would mask the blood flow signal.

What would settle it

Running the system on the ExpDrive dataset but replacing the predictive model with a fixed exposure setting and observing whether the performance improvement disappears would test if the adaptation step is necessary.

Figures

Figures reproduced from arXiv: 2605.04397 by Caifeng Shan, Jieying Wang, Wenjin Wang, Xinqi Cai.

**Figure 1.** Figure 1: Conceptual overview of the proposed “Optimize-at-Capture” framework in dynamic driving environments. The top row (red path) illustrates signal view at source ↗

**Figure 2.** Figure 2: Characteristic curves showing the relationship between exposure time view at source ↗

**Figure 3.** Figure 3: The core premise is that within a short temporal window and a stable scene, the relationship between exposure time T and the average facial pixel intensity I remains linear: I = k · T + b. (7) Our algorithm dynamically estimates the parameters k (slope, reflecting the influence of exposure on brightness) and b (intercept) to adapt to the changing illumination view at source ↗

**Figure 3.** Figure 3: Triplet-frame adaptive exposure control flowchart. view at source ↗

**Figure 4.** Figure 4: Overview of the experimental design and in-vehicle setup. view at source ↗

**Figure 5.** Figure 5: Age distribution (left) and heart rate distribution (right) of subjects view at source ↗

**Figure 6.** Figure 6: Distributional comparison of rPPG performance across exposure strategies ( view at source ↗

**Figure 7.** Figure 7: Performance Comparison Under Different Weather Conditions. view at source ↗

**Figure 8.** Figure 8: Comparison of the performance of four exposure control methods view at source ↗

**Figure 9.** Figure 9: Qualitative performance comparison across four representative driving conditions. From top to bottom: Case 1 (typical conditions), Case 2 (extreme view at source ↗

**Figure 10.** Figure 10: Performance comparison and ablation study of the Multi-Exposure Region Fusion (MERF) module ( view at source ↗

read the original abstract

Remote photoplethysmography (rPPG) holds great promise for continuous heart-rate monitoring of drivers in intelligent vehicles. However, its performance is severely degraded by the highly dynamic illumination changes. A critical yet overlooked factor is the lack of exposure controlling during video acquisition -- most existing systems rely on either fixed exposure settings or camera build-in auto-exposure, both of which fail to maintain stable facial brightness under rapidly changing lighting conditions during driving. To address this gap, we propose a highly-adaptive exposure controlling framework that proactively adjusts exposure parameters based on predictive modeling of historical skin reflections. Unlike standard auto-exposure, our method is specifically optimized for rPPG measurement, ensuring the skin region of interest (ROI) remains within the optimal dynamic range for rPPG signal extraction. As an important contribution of this study, we introduce ExpDrive, a public in-vehicle physiological monitoring dataset comprising synchronized facial video and reference ECG from 48 subjects captured under real driving conditions. Extensive experiments demonstrate that our method consistently outperforms fixed exposure and standard auto-exposure strategies. Specifically, it reduces the Mean Absolute Error (MAE) by 6.31 bpm (from 14.1 to 7.79 bpm) and significantly increases the success rate by 32.3 percentage points (p < 0.001) (from 24.9% to 57.2%) across challenging driving scenarios. Notably, it clearly improved the performance of non-contact heart-rate monitoring in both low-light (rainy) and high-glare (sunny) conditions, validating the efficacy of exposure-aware acquisition design.

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

This paper offers a practical exposure control method tuned for rPPG in cars plus a new real-driving dataset, but the reported gains rest on a predictive step whose side effects are not yet shown in detail.

read the letter

The main thing to know is that this work targets camera exposure settings at capture time to stabilize skin brightness for remote heart rate monitoring while driving, and it includes a new public dataset collected in actual vehicles. They move away from fixed exposure or the camera's built-in auto mode by using a predictive model based on past skin reflections to choose parameters that keep the ROI in the right intensity range for rPPG extraction. That is a reasonable shift because rPPG depends on small, consistent color changes rather than general image appeal. Releasing ExpDrive with 48 subjects, synchronized videos, and ECG under real conditions fills a gap, since most existing sets do not capture the rapid lighting shifts from sun, rain, or road movement. The numbers they give show a drop in mean absolute error from 14.1 bpm to 7.79 bpm and success rate rising from 24.9% to 57.2%, with statistical significance, and they note better results in both low-light and high-glare cases. That is concrete evidence of improvement on their test set. The soft spot is the predictive modeling itself. If the forecast from historical data is off even modestly in fast-changing light, the exposure adjustments could create brightness jumps or clipping that add non-physiological variance to the pixel series. Standard filtering may not remove all of it, yet the abstract gives no per-condition predictor error rates, no SNR measurements tied to the exposure changes, and no ablation on whether smoothing the transitions preserves the pulsatile signal. Without those checks, it is hard to separate the benefit of better average exposure from possible artifacts introduced by the control loop. This paper is for applied researchers building driver monitoring systems or anyone extending rPPG to uncontrolled environments. The dataset alone makes it worth examining for that group. It deserves peer review because the problem is real, the dataset is new, and the targeted optimization is a clear step, even though the central claim would need tighter validation on the prediction reliability and noise analysis.

Referee Report

2 major / 1 minor

Summary. The manuscript proposes a predictive exposure control framework for remote photoplethysmography (rPPG) heart-rate monitoring in vehicles. It adjusts camera exposure in real time by modeling historical skin reflections to keep the facial ROI within the optimal dynamic range, unlike fixed or standard auto-exposure. The authors release the ExpDrive dataset (48 subjects, synchronized facial video and ECG under real driving) and report that the method lowers MAE by 6.31 bpm (14.1 to 7.79 bpm) and raises success rate by 32.3 percentage points (24.9% to 57.2%, p<0.001) versus baselines, with gains in both low-light and high-glare conditions.

Significance. If the central claims are substantiated, the work addresses a practical bottleneck in in-vehicle rPPG by treating exposure as an optimizable acquisition parameter rather than a post-hoc correction. The public ExpDrive dataset is a clear strength that supports reproducibility and further benchmarking. The reported effect sizes are large and statistically tested on held-out driving data, which would be valuable for driver-monitoring applications if the adaptive control is shown not to inject non-physiological variance into the pulsatile signal.

major comments (2)

[Abstract] Abstract and method description: the performance gains (MAE drop of 6.31 bpm, success-rate increase of 32.3 pp) rest on the assumption that real-time predictive exposure adjustment preserves rPPG signal quality. No predictor error rates, SNR/ablation results on exposure-induced noise, or evidence that transitions are smoothed to avoid abrupt intensity jumps are supplied; without these, it is unclear whether the reported improvements could be offset by added temporal artifacts in rapidly varying illumination.
[Evaluation] Dataset and evaluation section: while ExpDrive is introduced as a public resource, the manuscript provides no per-condition breakdown of predictor accuracy or exposure-induced variance in the extracted rPPG traces. This information is load-bearing for the claim that the method is specifically optimized for rPPG rather than generic image quality.

minor comments (1)

[Abstract] The abstract states quantitative results but does not indicate the number of subjects or trials used for the statistical test (p<0.001); adding this detail would improve clarity.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive and detailed feedback. We have carefully reviewed the major comments and provide point-by-point responses below. We agree that additional analyses are needed to fully substantiate the claims and will revise the manuscript accordingly.

read point-by-point responses

Referee: [Abstract] Abstract and method description: the performance gains (MAE drop of 6.31 bpm, success-rate increase of 32.3 pp) rest on the assumption that real-time predictive exposure adjustment preserves rPPG signal quality. No predictor error rates, SNR/ablation results on exposure-induced noise, or evidence that transitions are smoothed to avoid abrupt intensity jumps are supplied; without these, it is unclear whether the reported improvements could be offset by added temporal artifacts in rapidly varying illumination.

Authors: We agree that the manuscript would be strengthened by explicit evidence that the predictive exposure control preserves (and does not degrade) rPPG signal quality. In the revised version we will (1) report the exposure predictor's accuracy (MAE and RMSE between predicted and target exposure values on held-out sequences), (2) add SNR and frequency-domain ablations comparing rPPG traces obtained under our method versus fixed and auto-exposure baselines, and (3) describe and illustrate the temporal smoothing applied to exposure transitions to eliminate abrupt intensity jumps. These additions will directly address whether any new temporal artifacts offset the reported gains. revision: yes
Referee: [Evaluation] Dataset and evaluation section: while ExpDrive is introduced as a public resource, the manuscript provides no per-condition breakdown of predictor accuracy or exposure-induced variance in the extracted rPPG traces. This information is load-bearing for the claim that the method is specifically optimized for rPPG rather than generic image quality.

Authors: We concur that per-condition breakdowns are necessary to demonstrate rPPG-specific optimization. In the revision we will add tables and figures that report, for each illumination regime (low-light/rainy, high-glare/sunny, and mixed), the exposure predictor error and the resulting rPPG trace statistics (SNR, pulsatile power, and variance). Because the ExpDrive dataset is already public, these supplementary analyses can be independently verified and will clarify that the method targets physiological signal fidelity rather than generic image quality metrics. revision: yes

Circularity Check

0 steps flagged

No circularity: empirical performance measured on held-out driving data against external baselines

full rationale

The paper proposes an exposure-control framework using predictive modeling of historical skin reflections, introduces the ExpDrive dataset, and reports MAE and success-rate improvements versus fixed-exposure and auto-exposure baselines. These metrics are obtained by applying the method to real driving videos and comparing against reference ECG; they are not obtained by fitting parameters to the target rPPG error or by re-using the same quantity as both input and output. No self-citation is invoked to justify a uniqueness theorem or to smuggle an ansatz, and the central claims rest on external, falsifiable measurements rather than definitional equivalence. The derivation chain therefore remains self-contained.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Central claim rests on the unstated premise that historical skin-reflection statistics are sufficient to predict future optimal exposure without additional domain-specific assumptions or fitted constants beyond those implicit in the predictive model.

pith-pipeline@v0.9.0 · 5605 in / 1187 out tokens · 28417 ms · 2026-05-08T18:07:03.241458+00:00 · methodology

discussion (0)

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

IndisputableMonolith.Cost (Jcost = ½(x+x⁻¹)−1) washburn_uniqueness_aczel unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

I = k·T + b ... k = (I_h − I_l)/(T_h − T_l), b = I_l − k·T_l ... T_opt = (I_target − b)/k

What do these tags mean?

matches: The paper's claim is directly supported by a theorem in the formal canon.
supports: The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends: The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses: The paper appears to rely on the theorem as machinery.
contradicts: The paper's claim conflicts with a theorem or certificate in the canon.
unclear: Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

37 extracted references · 37 canonical work pages

[1]

W.-Y . C. Sang-Joong Jung, Heung-Sub Shin, Driver fatigue and drowsi- ness monitoring system with embedded electrocardiogram sensor on steering wheel, IET Intelligent Transport Systems 8 (2014) 43–50

work page 2014
[2]

Matsuda, M

T. Matsuda, M. Makikawa, ECG monitoring of a car driver using capacitively-coupled electrodes, in: 2008 30th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, 2008, pp. 1315–1318

work page 2008
[3]

Leicht, P

L. Leicht, P. Vetter, S. Leonhardt, D. Teichmann, The physiobelt: A safety belt integrated sensor system for heart activity and respiration, in: 2017 IEEE International Conference on Vehicular Electronics and Safety (ICVES), 2017, pp. 191–195

work page 2017
[4]

Huang, S

B. Huang, S. Hu, Z. Liu, C.-L. Lin, J. Su, C. Zhao, L. Wang, W. Wang, Challenges and prospects of visual contactless physiological monitoring in clinical study, NPJ Digital Medicine 6 (2023)

work page 2023
[5]

J. Wang, C. Shan, L. Liu, Z. Hou, Camera-based physiological mea- surement: Recent advances and future prospects, Neurocomputing 575 (2024) 127282

work page 2024
[6]

Leonhardt, L

S. Leonhardt, L. Leicht, D. Teichmann, Unobtrusive vital sign monitor- ing in automotive environments—a review, Sensors 18 (9) (2018). URL https://www.mdpi.com/1424-8220/18/9/3080

work page 2018
[7]

E. M. Nowara, T. K. Marks, H. Mansour, A. Veeraraghavan, Near- infrared imaging photoplethysmography during driving, IEEE Transac- tions on Intelligent Transportation Systems 23 (4) (2022) 3589–3600

work page 2022
[8]

Hernandez-Ortega, S

J. Hernandez-Ortega, S. Nagae, J. Fierrez, A. Morales, Quality-based pulse estimation from nir face video with application to driver monitor- ing, Springer, Cham (2019)

work page 2019
[9]

E. M. Nowara, T. K. Marks, H. Mansour, A. Veeraraghavan, Sparseppg: Towards driver monitoring using camera-based vital signs estimation in near-infrared, in: 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), 2018, pp. 1385–1394

work page 2018
[10]

Comas, T

A. Comas, T. K. Marks, H. Mansour, S. Lohit, Y . Ma, X. Liu, Turnip: Time-series u-net with recurrence for nir imaging ppg, in: 2021 IEEE International Conference on Image Processing (ICIP), 2021, pp. 309– 313

work page 2021
[11]

Z. Gong, X. Yang, R. Song, X. Han, C. Ren, H. Shi, J. Niu, W. Li, Heart rate estimation in driver monitoring system using quality-guided spectrum peak screening, IEEE Transactions on Instrumentation and Measurement 73 (2024) 1–14

work page 2024
[12]

Vizbara, Comparison of green, blue and infrared light in wrist and forehead photoplethysmography, in: BioMed, 2013

V . Vizbara, Comparison of green, blue and infrared light in wrist and forehead photoplethysmography, in: BioMed, 2013. URL https://api.semanticscholar.org/CorpusID:110981655

work page 2013
[13]

Corral-Martinez, G

L. Corral-Martinez, G. Paez, M. Strojnik, Optimal wavelength selection for noncontact reflection photoplethysmography, Proceedings of SPIE - The International Society for Optical Engineering 8011 (2011) 304–. doi:10.1117/12.903190

work page doi:10.1117/12.903190 2011
[14]

L. Chiu, Y . Chou, Y . Wu, B. Wu, Deep-learning-based remote pho- toplethysmography measurement in driving scenarios with color and near-infrared images, IEEE Transactions on Instrumentation and Mea- surement 72 (2023) 1–12

work page 2023
[15]

Park, B.-K

S. Park, B.-K. Kim, S.-Y . Dong, Self-supervised rgb-nir fusion video vision transformer framework for rppg estimation, IEEE Transactions on Instrumentation and Measurement 71 (2022) 1–10

work page 2022
[16]

J. Wang, X. Yang, Q. Hu, J. Tang, C. Liu, D. He, Y . Wang, Y . Chen, K. Wu, Physdrive: A multimodal remote physiological measurement dataset for in-vehicle driver monitoring (2025). arXiv:2507.19172. URL https://arxiv.org/abs/2507.19172

work page arXiv 2025
[17]

Wu, Y .-W

B.-F. Wu, Y .-W. Chu, P.-W. Huang, M.-L. Chung, Neural network based luminance variation resistant remote-photoplethysmography for driver’s heart rate monitoring, IEEE Access 7 (2019) 57210–57225

work page 2019
[18]

Wu, Y .-C

B.-F. Wu, Y .-C. Wu, Y .-W. Chou, A compensation network with er- ror mapping for robust remote photoplethysmography in noise-heavy conditions, IEEE Transactions on Instrumentation and Measurement 71 (2022) 1–11

work page 2022
[19]

Huang, B.-J

P.-W. Huang, B.-J. Wu, B.-F. Wu, A heart rate monitoring framework for real-world drivers using remote photoplethysmography, IEEE Journal of Biomedical and Health Informatics 25 (5) (2021) 1397–1408

work page 2021
[20]

Jiang, K.-D

T. Jiang, K.-D. Kuhnert, D. Nguyen, L. Kuhnert, Multiple templates auto exposure control based on luminance histogram for onboard camera, in: 2011 IEEE International Conference on Computer Science and Automation Engineering, V ol. 3, 2011, pp. 237–241

work page 2011
[21]

Schulz, M

S. Schulz, M. Grimm, R.-R. Grigat, Using brightness histogram to perform optimum auto exposure, WSEAS Trans. System and Control 2 (2007) 123–128

work page 2007
[22]

Q. K. Vuong, S.-H. Yun, S. Kim, A new auto exposure system to detect high dynamic range conditions using cmos technology, in: 2008 Third International Conference on Convergence and Hybrid Information Technology, V ol. 1, 2008, pp. 577–580

work page 2008
[23]

H.-W. Kim, S. Kwon, J. Je-Kyo, J. ha, Auto-exposure control method for a stereo camera robust to brightness variation, International Journal of Control and Automation 7 (2014) 321–330. JOURNAL OF LATEX CLASS FILES, VOL. 18, NO. 9, SEPTEMBER 2020 13

work page 2014
[24]

S. Yi, D. Yu, Y . Zhu, W. Wang, Illumination-robust camera-ppg via skin- guided auto-exposure, in: 2024 46th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), 2024, pp. 1–4

work page 2024
[25]

Odinaev, J

I. Odinaev, J. W. Chin, K. Ho Luo, Z. Ke, R. H. SO, K. Long Wong, Optimizing camera exposure control settings for remote vital sign mea- surements in low-light environments, in: 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), 2023, pp. 6086–6093

work page 2023
[26]

J. M. Warnecke, J. Lasenby, T. M. Deserno, Robust in-vehicle heartbeat detection using multimodal signal fusion, Scientific Reports 13 (2023)

work page 2023
[27]

Othman, A

W. Othman, A. Kashevnik, A. Ali, N. Shilov, Drivermvt: In-cabin dataset for driver monitoring including video and vehicle telemetry information, Data 7 (5) (2022)

work page 2022
[28]

Xiang, S

G. Xiang, S. Yao, H. Deng, X. Wu, X. Wang, Q. Xu, T. Yu, K. Wang, Y . Peng, A multi-modal driver emotion dataset and study: Including facial expressions and synchronized physiological signals, Engineering Applications of Artificial Intelligence 130 (2024) 107772

work page 2024
[29]

M. Xu, G. Zeng, Y . Song, Y . Cao, Z. Liu, X. He, Ivrr-ppg: An illumination variation robust remote-ppg algorithm for monitoring heart rate of drivers, IEEE Transactions on Instrumentation and Measurement 72 (2023) 1–10

work page 2023
[30]

W. Liu, X. Fang, Q. Chen, Y . Li, T. Li, Reliability analysis of an integrated device of ecg, ppg and pressure pulse wave for cardiovascular disease, Microelectronics Reliability 87 (AUG.) (2018) 183–187

work page 2018
[31]

S. Hu, Q. Gao, K. Xie, C. Wen, W. Zhang, J. He, Efficient detection of driver fatigue state based on all-weather illumination scenarios, Scientific Reports 14 (17075) (2024) 17075

work page 2024
[32]

K. H. Ang, G. Chong, Y . Li, Pid control system analysis, design, and technology, IEEE Transactions on Control Systems Technology 13 (4) (2005) 559–576

work page 2005
[33]

Laurie, N

J. Laurie, N. Higgins, T. Peynot, J. Roberts, Dedicated exposure control for remote photoplethysmography, IEEE Access 8 (2020) 116642– 116652

work page 2020
[34]

W. Wang, A. C. den Brinker, S. Stuijk, G. de Haan, Algorithmic principles of remote PPG, IEEE T BIO-MED ENG 64 (7) (2017) 1479– 1491

work page 2017
[35]

W. Wang, A. C. den Brinker, G. de Haan, Discriminative signatures for remote-ppg, IEEE Transactions on Biomedical Engineering 67 (2020) 1462–1473

work page 2020
[36]

W. Chen, R. Cordero, J. Taylor, D. Pangallo, R. Picard, M. Cruz, G. Regalia, Multicenter evaluation of machine-learning continuous pulse rate algorithm on wrist-worn device, Digital Biomarkers 8 (2024) 218– 228

work page 2024
[37]

Pianosi, T

F. Pianosi, T. Wagener, A simple and efficient method for global sensi- tivity analysis based on cumulative distribution functions, Environmental Modelling & Software 67 (2015) 1–11. Jieying Wangreceived the B.E. degree in elec- tronic and information engineering from China Jil- iang University, China, in 2015, and the M.S. in marine information science a...

work page 2015

[1] [1]

W.-Y . C. Sang-Joong Jung, Heung-Sub Shin, Driver fatigue and drowsi- ness monitoring system with embedded electrocardiogram sensor on steering wheel, IET Intelligent Transport Systems 8 (2014) 43–50

work page 2014

[2] [2]

Matsuda, M

T. Matsuda, M. Makikawa, ECG monitoring of a car driver using capacitively-coupled electrodes, in: 2008 30th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, 2008, pp. 1315–1318

work page 2008

[3] [3]

Leicht, P

L. Leicht, P. Vetter, S. Leonhardt, D. Teichmann, The physiobelt: A safety belt integrated sensor system for heart activity and respiration, in: 2017 IEEE International Conference on Vehicular Electronics and Safety (ICVES), 2017, pp. 191–195

work page 2017

[4] [4]

Huang, S

B. Huang, S. Hu, Z. Liu, C.-L. Lin, J. Su, C. Zhao, L. Wang, W. Wang, Challenges and prospects of visual contactless physiological monitoring in clinical study, NPJ Digital Medicine 6 (2023)

work page 2023

[5] [5]

J. Wang, C. Shan, L. Liu, Z. Hou, Camera-based physiological mea- surement: Recent advances and future prospects, Neurocomputing 575 (2024) 127282

work page 2024

[6] [6]

Leonhardt, L

S. Leonhardt, L. Leicht, D. Teichmann, Unobtrusive vital sign monitor- ing in automotive environments—a review, Sensors 18 (9) (2018). URL https://www.mdpi.com/1424-8220/18/9/3080

work page 2018

[7] [7]

E. M. Nowara, T. K. Marks, H. Mansour, A. Veeraraghavan, Near- infrared imaging photoplethysmography during driving, IEEE Transac- tions on Intelligent Transportation Systems 23 (4) (2022) 3589–3600

work page 2022

[8] [8]

Hernandez-Ortega, S

J. Hernandez-Ortega, S. Nagae, J. Fierrez, A. Morales, Quality-based pulse estimation from nir face video with application to driver monitor- ing, Springer, Cham (2019)

work page 2019

[9] [9]

E. M. Nowara, T. K. Marks, H. Mansour, A. Veeraraghavan, Sparseppg: Towards driver monitoring using camera-based vital signs estimation in near-infrared, in: 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), 2018, pp. 1385–1394

work page 2018

[10] [10]

Comas, T

A. Comas, T. K. Marks, H. Mansour, S. Lohit, Y . Ma, X. Liu, Turnip: Time-series u-net with recurrence for nir imaging ppg, in: 2021 IEEE International Conference on Image Processing (ICIP), 2021, pp. 309– 313

work page 2021

[11] [11]

Z. Gong, X. Yang, R. Song, X. Han, C. Ren, H. Shi, J. Niu, W. Li, Heart rate estimation in driver monitoring system using quality-guided spectrum peak screening, IEEE Transactions on Instrumentation and Measurement 73 (2024) 1–14

work page 2024

[12] [12]

Vizbara, Comparison of green, blue and infrared light in wrist and forehead photoplethysmography, in: BioMed, 2013

V . Vizbara, Comparison of green, blue and infrared light in wrist and forehead photoplethysmography, in: BioMed, 2013. URL https://api.semanticscholar.org/CorpusID:110981655

work page 2013

[13] [13]

Corral-Martinez, G

L. Corral-Martinez, G. Paez, M. Strojnik, Optimal wavelength selection for noncontact reflection photoplethysmography, Proceedings of SPIE - The International Society for Optical Engineering 8011 (2011) 304–. doi:10.1117/12.903190

work page doi:10.1117/12.903190 2011

[14] [14]

L. Chiu, Y . Chou, Y . Wu, B. Wu, Deep-learning-based remote pho- toplethysmography measurement in driving scenarios with color and near-infrared images, IEEE Transactions on Instrumentation and Mea- surement 72 (2023) 1–12

work page 2023

[15] [15]

Park, B.-K

S. Park, B.-K. Kim, S.-Y . Dong, Self-supervised rgb-nir fusion video vision transformer framework for rppg estimation, IEEE Transactions on Instrumentation and Measurement 71 (2022) 1–10

work page 2022

[16] [16]

J. Wang, X. Yang, Q. Hu, J. Tang, C. Liu, D. He, Y . Wang, Y . Chen, K. Wu, Physdrive: A multimodal remote physiological measurement dataset for in-vehicle driver monitoring (2025). arXiv:2507.19172. URL https://arxiv.org/abs/2507.19172

work page arXiv 2025

[17] [17]

Wu, Y .-W

B.-F. Wu, Y .-W. Chu, P.-W. Huang, M.-L. Chung, Neural network based luminance variation resistant remote-photoplethysmography for driver’s heart rate monitoring, IEEE Access 7 (2019) 57210–57225

work page 2019

[18] [18]

Wu, Y .-C

B.-F. Wu, Y .-C. Wu, Y .-W. Chou, A compensation network with er- ror mapping for robust remote photoplethysmography in noise-heavy conditions, IEEE Transactions on Instrumentation and Measurement 71 (2022) 1–11

work page 2022

[19] [19]

Huang, B.-J

P.-W. Huang, B.-J. Wu, B.-F. Wu, A heart rate monitoring framework for real-world drivers using remote photoplethysmography, IEEE Journal of Biomedical and Health Informatics 25 (5) (2021) 1397–1408

work page 2021

[20] [20]

Jiang, K.-D

T. Jiang, K.-D. Kuhnert, D. Nguyen, L. Kuhnert, Multiple templates auto exposure control based on luminance histogram for onboard camera, in: 2011 IEEE International Conference on Computer Science and Automation Engineering, V ol. 3, 2011, pp. 237–241

work page 2011

[21] [21]

Schulz, M

S. Schulz, M. Grimm, R.-R. Grigat, Using brightness histogram to perform optimum auto exposure, WSEAS Trans. System and Control 2 (2007) 123–128

work page 2007

[22] [22]

Q. K. Vuong, S.-H. Yun, S. Kim, A new auto exposure system to detect high dynamic range conditions using cmos technology, in: 2008 Third International Conference on Convergence and Hybrid Information Technology, V ol. 1, 2008, pp. 577–580

work page 2008

[23] [23]

H.-W. Kim, S. Kwon, J. Je-Kyo, J. ha, Auto-exposure control method for a stereo camera robust to brightness variation, International Journal of Control and Automation 7 (2014) 321–330. JOURNAL OF LATEX CLASS FILES, VOL. 18, NO. 9, SEPTEMBER 2020 13

work page 2014

[24] [24]

S. Yi, D. Yu, Y . Zhu, W. Wang, Illumination-robust camera-ppg via skin- guided auto-exposure, in: 2024 46th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), 2024, pp. 1–4

work page 2024

[25] [25]

Odinaev, J

I. Odinaev, J. W. Chin, K. Ho Luo, Z. Ke, R. H. SO, K. Long Wong, Optimizing camera exposure control settings for remote vital sign mea- surements in low-light environments, in: 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), 2023, pp. 6086–6093

work page 2023

[26] [26]

J. M. Warnecke, J. Lasenby, T. M. Deserno, Robust in-vehicle heartbeat detection using multimodal signal fusion, Scientific Reports 13 (2023)

work page 2023

[27] [27]

Othman, A

W. Othman, A. Kashevnik, A. Ali, N. Shilov, Drivermvt: In-cabin dataset for driver monitoring including video and vehicle telemetry information, Data 7 (5) (2022)

work page 2022

[28] [28]

Xiang, S

G. Xiang, S. Yao, H. Deng, X. Wu, X. Wang, Q. Xu, T. Yu, K. Wang, Y . Peng, A multi-modal driver emotion dataset and study: Including facial expressions and synchronized physiological signals, Engineering Applications of Artificial Intelligence 130 (2024) 107772

work page 2024

[29] [29]

M. Xu, G. Zeng, Y . Song, Y . Cao, Z. Liu, X. He, Ivrr-ppg: An illumination variation robust remote-ppg algorithm for monitoring heart rate of drivers, IEEE Transactions on Instrumentation and Measurement 72 (2023) 1–10

work page 2023

[30] [30]

W. Liu, X. Fang, Q. Chen, Y . Li, T. Li, Reliability analysis of an integrated device of ecg, ppg and pressure pulse wave for cardiovascular disease, Microelectronics Reliability 87 (AUG.) (2018) 183–187

work page 2018

[31] [31]

S. Hu, Q. Gao, K. Xie, C. Wen, W. Zhang, J. He, Efficient detection of driver fatigue state based on all-weather illumination scenarios, Scientific Reports 14 (17075) (2024) 17075

work page 2024

[32] [32]

K. H. Ang, G. Chong, Y . Li, Pid control system analysis, design, and technology, IEEE Transactions on Control Systems Technology 13 (4) (2005) 559–576

work page 2005

[33] [33]

Laurie, N

J. Laurie, N. Higgins, T. Peynot, J. Roberts, Dedicated exposure control for remote photoplethysmography, IEEE Access 8 (2020) 116642– 116652

work page 2020

[34] [34]

W. Wang, A. C. den Brinker, S. Stuijk, G. de Haan, Algorithmic principles of remote PPG, IEEE T BIO-MED ENG 64 (7) (2017) 1479– 1491

work page 2017

[35] [35]

W. Wang, A. C. den Brinker, G. de Haan, Discriminative signatures for remote-ppg, IEEE Transactions on Biomedical Engineering 67 (2020) 1462–1473

work page 2020

[36] [36]

W. Chen, R. Cordero, J. Taylor, D. Pangallo, R. Picard, M. Cruz, G. Regalia, Multicenter evaluation of machine-learning continuous pulse rate algorithm on wrist-worn device, Digital Biomarkers 8 (2024) 218– 228

work page 2024

[37] [37]

Pianosi, T

F. Pianosi, T. Wagener, A simple and efficient method for global sensi- tivity analysis based on cumulative distribution functions, Environmental Modelling & Software 67 (2015) 1–11. Jieying Wangreceived the B.E. degree in elec- tronic and information engineering from China Jil- iang University, China, in 2015, and the M.S. in marine information science a...

work page 2015