[Paper Review] VitalVideos-Europe: A dataset of face videos with PPG and blood pressure ground truths

Pieter-Jan Toye|arXiv (Cornell University)|Jun 2, 2023

Non-Invasive Vital Sign MonitoringEngineering3 citations

TL;DR

VitalVideos-Europe is a large-scale dataset of 850 participants featuring synchronized face videos, photoplethysmography (PPG) waveforms, and blood pressure measurements. Collected across diverse European locations with varied lighting and backgrounds, it supports research in remote vital sign estimation using video-based PPG and blood pressure validation, with demographic metadata including age, gender, and skin color.

ABSTRACT

We collected a large dataset consisting of 850 unique participants. For every participant we recorded two 30 second uncompressed videos, synchronized PPG waveforms and a single blood pressure measurement. Gender, age and skin color were also registered for every participant. The dataset includes roughly equal numbers of males and females, as well as participants of all ages. While the skin color distribution could have been more balanced, the dataset contains individuals from every skin color. The data was collected in a diverse set of locations to ensure a wide variety of backgrounds and lighting conditions. In an effort to assist in the research and development of remote vital sign measurement we are now opening up access to this dataset. vitalvideos.org for all datasets.

Motivation & Objective

• To create a large, diverse dataset for training and evaluating remote photoplethysmography (rPPG) and blood pressure estimation algorithms.
• To ensure representation across gender, age, and skin color to improve model generalization and fairness.
• To collect data under real-world conditions with varied lighting and backgrounds to enhance robustness of remote sensing methods.
• To provide a publicly accessible benchmark dataset with high-quality ground truth for PPG and blood pressure.

Proposed method

• Data collection involved recording two 30-second uncompressed face videos per participant using standard video equipment.
• Synchronized PPG waveforms were acquired using a finger-clip sensor to provide physiological ground truth.
• A single blood pressure measurement was recorded per participant using clinical sphygmomanometry.
• Demographic information—gender, age, and skin color—was collected for each participant to enable subgroup analysis.
• Participants were recruited across multiple European locations to ensure environmental diversity in lighting and background conditions.
• The dataset was curated and released with metadata and access links to support reproducibility and research use.

Experimental results

Research questions

• RQ1 Can a large-scale, diverse dataset improve the generalization of remote PPG and blood pressure estimation models across different skin types and lighting conditions?
• RQ2 How does demographic diversity—particularly in skin color and age—affect the performance of video-based vital sign estimation algorithms?
• RQ3 To what extent do real-world environmental variations in lighting and background impact the accuracy of rPPG and blood pressure estimation?
• RQ4 Can a publicly available dataset with ground-truth PPG and blood pressure measurements accelerate research in non-contact vital sign monitoring?

Key findings

• The dataset comprises 850 unique participants with balanced representation across male and female genders and a broad age range.
• Participants include individuals of all skin color categories, enhancing the dataset's diversity for fairness-aware algorithm development.
• The data was collected in real-world environments across Europe, resulting in significant variation in lighting and background conditions.
• Two 30-second uncompressed face videos were recorded per participant, providing high temporal resolution for signal analysis.
• Synchronized PPG waveforms and a single blood pressure measurement were acquired per participant, enabling ground-truth validation.
• The dataset is publicly released with full metadata, supporting reproducible research in remote vital sign monitoring.