kyegorov/mcd_rppg

Dataset Card for Dataset Name

MCD-rPPG (Multi-Camera Dataset for Remote Photoplethysmography) is a large-scale multimodal dataset designed for remote photoplethysmography (rPPG) and health biomarker estimation from facial video.

Dataset Details

Dataset Description

MCD-rPPG (Multi-Camera Dataset for Remote Photoplethysmography) is a large-scale multimodal dataset designed for remote photoplethysmography (rPPG) and health biomarker estimation from facial video. The dataset addresses critical limitations of existing rPPG datasets, including small size, privacy concerns with facial videos, and lack of diversity in recording conditions.

The dataset comprises 3,600 synchronized video recordings from 600 subjects, captured under varied conditions (resting and post-exercise) using multiple consumer-grade cameras at different angles. Each recording is paired with 100 Hz PPG and ECG signals, along with 13 extended health biomarkers.

• License: cc-by-4.0

Dataset Sources

Repository: https://github.com/ksyegorov/mcd_rppg

Hugging Face Hub: https://huggingface.co/datasets/kyegorov/mcd_rppg

Paper: Egorov, K., Botman, S., Blinov, P., Zubkova, G., Ivaschenko, A., Kolsanov, A., & Savchenko, A. (2025). Gaze into the Heart: A Multi-View Video Dataset for rPPG and Health Biomarkers Estimation. Proceedings of the 33rd ACM International Conference on Multimedia. DOI: 10.1145/3746027.3758255

arXiv: https://arxiv.org/abs/2508.17924v1

Uses

Direct Use

The dataset is intended for:

• Training and evaluating rPPG models for pulse wave signal estimation from facial video

• Multi-task learning for simultaneous estimation of PPG signals and health biomarkers

• Cross-dataset evaluation of rPPG model generalization

• Research on contactless health monitoring and telemedicine applications

Out-of-Scope Use

Clinical diagnosis or medical decision-making without validation

• Identification or re-identification of subjects from facial video data

• Commercial deployment without proper validation and ethical review

• Use of facial videos for biometric recognition or surveillance purposes

Dataset Structure

The dataset is organized as follows:

3600 video recordings (600 subjects × 2 states × 3 cameras)

Synchronized PPG (100 Hz) and ECG signals

13 health biomarkers: systolic/diastolic pressure, oxygen saturation, temperature, glucose, glycated hemoglobin, cholesterol, respiratory rate, arterial stiffness, stress level (PSM-25), age, sex, BMI

Multi-view videos: frontal webcam, FullHD camcorder, mobile phone camera

Each subject was recorded twice: before and after physical activity (e.g., 15 squats in 30 seconds). Each measurement corresponds to a three-minute video simultaneously captured from three different angles.

Dataset Creation

Curation Rationale

The dataset was created to address the critical limitations of existing publicly available rPPG datasets:

• Small size: Existing datasets typically contain only 10–140 subjects

• Privacy concerns: Most known datasets require restricted access with no guarantee of approval, impeding collaborative research

• Lack of diversity: Insufficient variation in recording conditions, subjects, and physiological states

• Limited health metrics: Most datasets lack comprehensive health data such as temperature, ECG, and other biomarkers

The public release of the dataset aims to significantly accelerate progress in the development of AI medical assistants.

Source Data

Data Collection and Processing

Data was collected using a multimodal capture setup involving three consumer-grade cameras at different angles: a frontal webcam, a FullHD camcorder, and a mobile phone camera. Each subject was recorded in two states: resting and post-exercise (e.g., 15 squats in 30 seconds).

Synchronized PPG signals were sampled at 100 Hz, alongside ECG signals. Extended health metrics including arterial blood pressure, oxygen saturation, temperature, glucose, glycated hemoglobin, cholesterol, respiratory rate, arterial stiffness, stress level (PSM-25), age, sex, and BMI were collected for each subject.

Personal and Sensitive Information

The dataset contains facial video recordings and personal health data, including age, sex, BMI, blood pressure, and other physiological measurements. These constitute personal and sensitive information. The dataset is released under CC BY 4.0, indicating that the data has been de-identified or anonymized to the extent possible. However, users should be aware that facial video data may still pose privacy risks, and the dataset should be used in accordance with applicable data protection regulations.

Bias, Risks, and Limitations

Demographic bias: The dataset may not be representative of global populations, as data was collected primarily in Russia

Recording conditions: While the dataset includes resting and post-exercise states, it may not cover all real-world scenarios (e.g., different lighting conditions, motion artifacts)

Privacy risks: Despite CC BY 4.0 licensing, facial video data carries inherent privacy risks

Clinical validity: The health biomarkers provided are research-grade measurements and should not be used for clinical diagnosis without proper validation

Model generalization: rPPG models trained on this dataset may not generalize well to other populations or recording conditions

Recommendations

Users should be aware of the risks, biases, and limitations of the dataset. More information is needed for further recommendations regarding:

• Fairness evaluation across demographic groups

• Robustness testing under varied environmental conditions

• Ethical use guidelines for facial video data

• Validation protocols for clinical applications

Citation

BibTeX:

@inproceedings{10.1145/3746027.3758255,
author = {Egorov, Konstantin and Botman, Stepan and Blinov, Pavel and Zubkova, Galina and Ivaschenko, Anton and Kolsanov, Alexander and Savchenko, Andrey},
title = {Gaze into the Heart: A Multi-View Video Dataset for rPPG and Health Biomarkers Estimation},
year = {2025},
isbn = {979-8-4007-2135-7},
publisher = {Association for Computing Machinery},
address = {New York, NY, USA},
url = {https://doi.org/10.1145/3746027.3758255},
doi = {10.1145/3746027.3758255},
booktitle = {Proceedings of the 33rd ACM International Conference on Multimedia},
pages = {},
numpages = {},
location = {Dublin, Ireland},
series = {MM '25}
}

APA:

Egorov, K., Botman, S., Blinov, P., Zubkova, G., Ivaschenko, A., Kolsanov, A., & Savchenko, A. (2025). Gaze into the Heart: A Multi-View Video Dataset for rPPG and Health Biomarkers Estimation. Proceedings of the 33rd ACM International Conference on Multimedia. https://doi.org/10.1145/3746027.3758255

Glossary

rPPG (Remote Photoplethysmography): A non-invasive technique for estimating pulse wave signals from video recordings of the skin

PPG (Photoplethysmography): An optical technique for detecting blood volume changes in microvascular tissue

ECG (Electrocardiogram): A recording of the electrical activity of the heart

PSM-25: Psychological Stress Measure, a 25-item self-report questionnaire for assessing stress level

ROI (Region of Interest): The facial region used for rPPG signal extraction