Camera-based HbA1c estimation using remote photoplethysmography: development, independent evaluation, and baseline comparator analysis
6 affiliations
- 1Department of Anaesthesiology, Singapore General Hospital, Singapore, Singapore
- 2Duke-NUS Medical School, Singapore, Singapore
- 3Nervotec, Singapore, Singapore
- 4Yong Loo Lin School of Medicine, National University of Singapore, Singapore, Singapore
- 5Bristol Medical School, University of Bristol, Bristol, United Kingdom
- 6Department of Endocrinology, Singapore General Hospital, Singapore, Singapore
Abstract
Type 2 diabetes mellitus is highly prevalent and frequently undiagnosed, yet HbA1c testing depends on blood sampling and laboratory infrastructure, limiting its use for low-friction dysglycemia screening. Remote photoplethysmography (rPPG) records facial color signals with a standard camera and has been proposed as a digital biomarker platform for cardiometabolic risk. We tested whether facial rPPG color time series carry information about laboratory HbA1c beyond simple comparators. In a prospective single-center study at Singapore General Hospital, 254 participants had both a usable facial recording and a laboratory HbA1c. Each recording was reduced to the mean red, green and blue values of six facial regions over 8.53 s. NervoHbA1c, a 4.27-million-parameter convolutional model with demographic inputs, was evaluated by pooled 5-fold cross-validation repeated three times, leave-one-cohort-out evaluation, a learning curve and testing on two independent smartphone cohorts, against mean, median and demographics-only comparators fitted within each training split. Within the development cohort, 16 models spanning 0–4.27 million fitted parameters were compared, including frozen-encoder linear read-outs and conventional signal-processing features. Pooled over 254 participants, NervoHbA1c had a mean absolute error (MAE) of 1.081% (95% CI 0.973–1.193), against 1.230% for the median baseline, 1.237% for demographics-only and 1.254% for the mean baseline (Nadeau–Bengio p = 0.008–0.042), and its predictions correlated with laboratory HbA1c (r = 0.505). It had the lowest error in every leave-one-cohort-out split, and its error fell steadily as the training set grew. In the development cohort, error decreased with model capacity: every lower-capacity model had higher error than the full model (MAE 0.732% against 0.844% for the mean baseline). A signal-only model was modestly correlated with HbA1c (r = 0.368). The published checkpoint, trained on 53 webcam participants, did not transfer to the smartphone cohorts (SGH-6 MAE 1.478% against 1.205% for demographics-only), a gap attributable to the change of acquisition device (webcam against smartphone AUROC 1.000). Short facial rPPG recordings therefore carry modest but recoverable information about HbA1c beyond demographics when the training data match the acquisition device. Participant-level evaluation, cross-cohort testing and simple comparators are needed before camera-based digital biomarkers advance toward clinical use, and a larger full-precision cohort is planned.
Related
- Stroboscopic Light Stimulation in Adults Reporting Depressive Symptoms: Safety, Tolerability, Feasibility, and Active-Comparator Development in a Staged Early-Phase Study
- Association of life course type 2 diabetes duration with adverse cardiac structure independent of current glycaemia in the British MRC National Survey of Health and Development Birth Cohort
- Development and External Validation of Machine Learning Models for Ionized Calcium Estimation in Hypoalbuminemia
- PedDx: An End-to-End Deep Learning Framework for Pediatric Posterior Fossa Tumor Diagnosis with Agentic AI-Enabled Independent External Evaluation
- A Reproducible Decomposed Large Language Model Pipeline for Structured Abstraction of Multimodal Neuroimaging Reports: Development and Independent Validation in 1,669 Emergency Department Encounters for Dizziness and Vertigo
BibTeX
@article{abdullah2026camerabased,
title = {{Camera-based HbA1c estimation using remote photoplethysmography: development, independent evaluation, and baseline comparator analysis}},
author = {Hairil Rizal Abdullah and Hairil Rizal Abdullah and Si Ying Tan and Isuru Kalhara and Shiqi Wang and Vera Jia Qi Lee and Roderica Rui Ge Ng and Roderica Rui Ge Ng and Yong Mong Bee and Yong Mong Bee},
journal = {Frontiers in Digital Health},
year = {2026},
month = nov,
doi = {10.3389/fdgth.2026.1978726},
url = {https://doi.org/10.3389/fdgth.2026.1978726},
note = {CC BY},
}
RIS
TY - JOUR
TI - Camera-based HbA1c estimation using remote photoplethysmography: development, independent evaluation, and baseline comparator analysis
AU - Hairil Rizal Abdullah
AU - Hairil Rizal Abdullah
AU - Si Ying Tan
AU - Isuru Kalhara
AU - Shiqi Wang
AU - Vera Jia Qi Lee
AU - Roderica Rui Ge Ng
AU - Roderica Rui Ge Ng
AU - Yong Mong Bee
AU - Yong Mong Bee
PY - 2026
DA - 2026/11/05
JO - Frontiers in Digital Health
DO - 10.3389/fdgth.2026.1978726
UR - https://doi.org/10.3389/fdgth.2026.1978726
AB - Type 2 diabetes mellitus is highly prevalent and frequently undiagnosed, yet HbA1c testing depends on blood sampling and laboratory infrastructure, limiting its use for low-friction dysglycemia screening. Remote photoplethysmography (rPPG) records facial color signals with a standard camera and has been proposed as a digital biomarker platform for cardiometabolic risk. We tested whether facial rPPG color time series carry information about laboratory HbA1c beyond simple comparators. In a prospective single-center study at Singapore General Hospital, 254 participants had both a usable facial recording and a laboratory HbA1c. Each recording was reduced to the mean red, green and blue values of six facial regions over 8.53 s. NervoHbA1c, a 4.27-million-parameter convolutional model with demographic inputs, was evaluated by pooled 5-fold cross-validation repeated three times, leave-one-cohort-out evaluation, a learning curve and testing on two independent smartphone cohorts, against mean, median and demographics-only comparators fitted within each training split. Within the development cohort, 16 models spanning 0–4.27 million fitted parameters were compared, including frozen-encoder linear read-outs and conventional signal-processing features. Pooled over 254 participants, NervoHbA1c had a mean absolute error (MAE) of 1.081% (95% CI 0.973–1.193), against 1.230% for the median baseline, 1.237% for demographics-only and 1.254% for the mean baseline (Nadeau–Bengio p = 0.008–0.042), and its predictions correlated with laboratory HbA1c (r = 0.505). It had the lowest error in every leave-one-cohort-out split, and its error fell steadily as the training set grew. In the development cohort, error decreased with model capacity: every lower-capacity model had higher error than the full model (MAE 0.732% against 0.844% for the mean baseline). A signal-only model was modestly correlated with HbA1c (r = 0.368). The published checkpoint, trained on 53 webcam participants, did not transfer to the smartphone cohorts (SGH-6 MAE 1.478% against 1.205% for demographics-only), a gap attributable to the change of acquisition device (webcam against smartphone AUROC 1.000). Short facial rPPG recordings therefore carry modest but recoverable information about HbA1c beyond demographics when the training data match the acquisition device. Participant-level evaluation, cross-cohort testing and simple comparators are needed before camera-based digital biomarkers advance toward clinical use, and a larger full-precision cohort is planned.
SN - 2673-253X
ER -